diff --git a/scripts/generate-index-for-system-varibles.py b/scripts/generate-index-for-system-varibles.py
new file mode 100644
index 0000000000000..e48a47e3226e0
--- /dev/null
+++ b/scripts/generate-index-for-system-varibles.py
@@ -0,0 +1,218 @@
+# This script enables you to update the "## Variable reference" section in `system-variable-reference.md` automatically according to the latest reference information of all system variables in TiDB documentation.
+# Before running this script, you need to specify the local directory of the TiDB source documentation in the `docs_dir` variable.
+# If a system variable name contains `_`, the script will add all references to this variable name to the reference index, except when the variable name is surrounded by `_` or `-` or other English letters in the referenced docs.
+# If a system variable name does not contain `_`, the script will only count the link reference of this variable as valid references and add the valid references to the reference index.
+
+import re
+from pathlib import Path
+import os
+
+docs_dir = "/Users/grcai/Documents/GitHub/docs-cn" # Specify the local directory of the TiDB source documentation, which can be either English or Chinese
+reference_section_titles = ["Variable reference", "变量索引"]
+referenced_in_text = ["Referenced in:", "引用该变量的文档："]
+
+def get_md_files_in_toc(toc_file):
+    with open(toc_file, 'r', encoding='utf-8') as f:
+        content = f.read()
+    return re.findall(r'\[.*?\]\(/(?:[^)]+/)?([^/]+\.md)(?:#[^)]*)?\)', content)
+
+def generate_var_link(variable_line):
+    # Remove <span> tags but keep the content inside span
+    variable_line = re.sub(r'\s*?<span.*?>(.*?)</span>\s*?', r' \1', variable_line).strip()
+    # Remove backticks
+    variable_line = re.sub(r'`', '', variable_line).strip()
+    variable_line = variable_line.lstrip("#").strip().lower() # Remove leading "#" and trim spaces, convert to lowercase
+    variable_line = re.sub(r"[^\w\u4e00-\u9fff\s-]", "", variable_line)  # Remove special symbols but keep spaces
+    variable_line = re.sub(r"\s+", "-", variable_line)  # Replace spaces with hyphens
+    return variable_line
+
+def extract_variables(content):
+    variables = []
+    for line in content.split('\n'):
+        if line.startswith('### '):
+            # Remove <span> tags
+            clean_line = re.sub(r'<span.*?</span>', '', line).strip()
+
+            # Extract variable name, match with or without backticks
+            var_name = re.sub(r'^### `(.*?)`|^### (.+)', r'\1\2', clean_line).strip()
+
+            if var_name:
+                var_link = generate_var_link(line)  # Variable name for generating links
+                variables.append((var_name, var_link))
+    return variables
+
+def find_references(variable_name, var_link, docs_dir, reference_paths_to_be_checked):
+    references = []
+    link_to_find = f"(/system-variables.md#{var_link})"
+
+    for path in reference_paths_to_be_checked:
+        try:
+            with open(path, 'r', encoding='utf-8') as f:
+                content = f.read()
+                if link_to_find in content:
+                        title = extract_doc_title(content, path)
+                        rel_path = str(path.relative_to(docs_dir))
+                        references.append((title, rel_path))
+                else:
+                    if "_" in variable_name and variable_name in content:
+                        if re.search(rf'(?<![A-Za-z_-]){re.escape(variable_name)}(?![A-Za-z_-])', content):
+                            title = extract_doc_title(content, path)
+                            rel_path = str(path.relative_to(docs_dir))
+                            references.append((title, rel_path))
+        except Exception as e:
+            print(f"Error reading file {path}: {e}")
+            continue
+
+    return references
+
+def extract_doc_title(content, path):
+    # Match level-1 heading that starts with # and has one or more spaces, trim trailing spaces
+    match = re.search(r'^#\s+(.*?)\s*$', content, re.MULTILINE)
+    if match:
+        return match.group(1)
+    return Path(path).stem
+
+def sort_references(references):
+    non_release_notes = []
+    release_notes = []
+
+    # Separate Release Notes from non-Release Notes
+    for title, path in references:
+        if path.startswith("releases/release-"):
+            release_notes.append((title, path))
+        else:
+            non_release_notes.append((title, path))
+
+    # Sort non-Release Notes by path title alphabetically
+    non_release_notes.sort(key=lambda x: x[0])
+
+    # Sort Release Notes by version number in descending order
+    def extract_version(path):
+        match = re.search(r'release-(\d+\.\d+(?:\.\d+)?)', path)
+        if match:
+            return tuple(map(int, match.group(1).split('.')))
+        return (0, 0, 0)  # Default version number as lowest
+
+    release_notes.sort(key=lambda x: extract_version(x[1]), reverse=True)
+
+    # Merge sorted results
+    return non_release_notes + release_notes
+
+def generate_reference_content(variables, docs_dir):
+    content = [f"## {reference_section_title}\n\n"]
+    unreferenced_variables = []  # Store unreferenced variables
+    reference_paths_to_be_checked = []
+    # Get reference_paths_to_be_checked list
+    for path in Path(docs_dir).rglob('*.md'):
+        if path.name in doc_md_list and path.name != reference_file and path.name != variables_file:
+            reference_paths_to_be_checked.append(path)
+
+    for var_name, var_link in sorted(variables):
+        refs = find_references(var_name, var_link, docs_dir, reference_paths_to_be_checked)
+        if not refs:
+            # Record variables that are not referenced by other documents
+            unreferenced_variables.append(var_name)
+        else:
+            pass
+
+        refs.append((variables_file_title, f"system-variables.md#{var_link}"))
+        sorted_refs = sort_references(refs) # Sort reference links
+
+        content.append(f"### {var_name}\n\n")
+        content.append(f"{referenced_in}\n\n")
+        for title, path in sorted_refs:
+            content.append(f"- [{title}](/{path})\n")
+        content.append("\n")
+
+    # Print unreferenced variables to console for reference
+    if unreferenced_variables:
+        print("\nUnreferenced Variables:")
+        print(f"There are {len(unreferenced_variables)} variables that are not referenced in any documents other than system-variables.md:")
+        for var in sorted(unreferenced_variables):
+            print(f"{var}")
+        print()
+
+    return ''.join(content)
+
+def update_reference_file_path(ref_file_path, new_content):
+    try:
+        with open(ref_file_path, 'r', encoding='utf-8') as f:
+            content = f.read()
+    except FileNotFoundError:
+        print(f"Warning: {ref_file_path} not found. Creating a new file.")
+        content = ""
+
+    start_pos = content.find(f'## {reference_section_title}')
+    if start_pos == -1:
+        print(f"Warning: '{reference_section_title}' section not found. Adding new section.")
+        updated_content = content + "\n" + new_content
+    else:
+        next_section_pos = content.find('\n## ', start_pos + 1)
+        if next_section_pos == -1:
+            next_section_pos = len(content)
+        updated_content = content[:start_pos] + new_content + content[next_section_pos:]
+
+    updated_content = re.sub(
+        rf'\[([^\]]+)\]\(/tidb-cloud/([^)]+)\.md\)',
+        rf'[\1]({cloud_docs_dir}/\2)',
+        updated_content
+    )
+
+    # Remove <span> tags
+    updated_content = re.sub(r'\[([^]]*?\s*?<span.*?</span>\s*?[^]]*)\]\(([^)]*)\)', lambda m: f'[{re.sub(r"\s*?<span.*?</span>\s*?", "", m.group(1))}]({m.group(2)})', updated_content)
+
+    updated_content = re.sub(r"What's New in TiDB 5.0", "TiDB 5.0 Release Notes", updated_content)
+
+    # Replace the last two empty lines with one empty line
+    updated_content = re.sub(r"\n\n$", "\n", updated_content)
+
+    with open(ref_file_path, 'w', encoding='utf-8') as f:
+        f.write(updated_content)
+
+def main():
+    global doc_md_list
+
+    if doc_language == "en":
+        doc_md_list = list(set(get_md_files_in_toc(tidb_file) + get_md_files_in_toc(tidb_cloud_file)))
+    else:
+        doc_md_list = list(set(get_md_files_in_toc(tidb_file)))
+
+    print("Start to generate reference information for system variables. It will take a few seconds...")
+
+    try:
+        # Read system variables
+        with open(variables_file_path, 'r', encoding='utf-8') as f:
+            variables = extract_variables(f.read())
+        # Generate new reference content
+        new_content = generate_reference_content(variables, docs_dir)
+        # Update reference file
+        update_reference_file_path(reference_file_path, new_content)
+        print("Reference file updated successfully!")
+    except Exception as e:
+        print(f"An error occurred: {e}")
+
+if __name__ == "__main__":
+
+    variables_file = "system-variables.md"
+    reference_file = "system-variable-reference.md"
+    tidb_toc_file = "TOC.md"
+    tidb_cloud_toc_file = "TOC-tidb-cloud.md"
+    cloud_docs_dir = "https://docs.pingcap.com/tidbcloud"
+    variables_file_path = os.path.abspath(os.path.join(docs_dir, variables_file))
+    reference_file_path = os.path.abspath(os.path.join(docs_dir, reference_file))
+    tidb_file = os.path.abspath(os.path.join(docs_dir, tidb_toc_file))
+    tidb_cloud_file = os.path.abspath(os.path.join(docs_dir, tidb_cloud_toc_file))
+
+    with open(variables_file_path, 'r', encoding='utf-8') as f:
+        variables_file_title = extract_doc_title(f.read(), variables_file_path)
+
+    if not re.search(r'[\u4e00-\u9fff]', variables_file_title): # Check if title contains no Chinese characters
+        doc_language = "en"
+        referenced_in = referenced_in_text[0]
+        reference_section_title = reference_section_titles[0]
+    else:
+        doc_language = "zh"
+        referenced_in = referenced_in_text[1]
+        reference_section_title = reference_section_titles[1]
+
+    main()