diff --git a/python/cudf/cudf/tests/test_array_ufunc.py b/python/cudf/cudf/tests/test_array_ufunc.py
index f1aad1af9e6..e4b4d5020ea 100644
--- a/python/cudf/cudf/tests/test_array_ufunc.py
+++ b/python/cudf/cudf/tests/test_array_ufunc.py
@@ -1,6 +1,8 @@
 # Copyright (c) 2020-2022, NVIDIA CORPORATION.
 
 import operator
+import warnings
+from contextlib import contextmanager
 from functools import reduce
 
 import cupy as cp
@@ -17,6 +19,37 @@
 ]
 
 
+@contextmanager
+def _hide_ufunc_warnings(ufunc):
+    # pandas raises warnings for some inputs to the following ufuncs:
+    name = ufunc.__name__
+    if name in {
+        "arccos",
+        "arccosh",
+        "arcsin",
+        "arctanh",
+        "fmod",
+        "log",
+        "log10",
+        "log2",
+        "reciprocal",
+    }:
+        with warnings.catch_warnings():
+            warnings.filterwarnings(
+                "ignore",
+                f"invalid value encountered in {name}",
+                category=RuntimeWarning,
+            )
+            warnings.filterwarnings(
+                "ignore",
+                f"divide by zero encountered in {name}",
+                category=RuntimeWarning,
+            )
+            yield
+    else:
+        yield
+
+
 @pytest.mark.parametrize("ufunc", _UFUNCS)
 @pytest.mark.parametrize("has_nulls", [True, False])
 @pytest.mark.parametrize("indexed", [True, False])
@@ -76,7 +109,8 @@ def test_ufunc_series(ufunc, has_nulls, indexed):
             pytest.xfail(reason="Operation not supported by cupy")
         raise
 
-    expect = ufunc(*(arg.to_pandas() for arg in pandas_args))
+    with _hide_ufunc_warnings(ufunc):
+        expect = ufunc(*(arg.to_pandas() for arg in pandas_args))
 
     try:
         if ufunc.nout > 1:
@@ -256,7 +290,8 @@ def test_ufunc_dataframe(ufunc, has_nulls, indexed):
             pytest.xfail(reason="Operation not supported by cupy")
         raise
 
-    expect = ufunc(*(arg.to_pandas() for arg in pandas_args))
+    with _hide_ufunc_warnings(ufunc):
+        expect = ufunc(*(arg.to_pandas() for arg in pandas_args))
 
     try:
         if ufunc.nout > 1: