diff --git a/docs/source/documents/api/agents/marl/mfac.rst b/docs/source/documents/api/agents/marl/mfac.rst
index d91d4410a..44634f966 100644
--- a/docs/source/documents/api/agents/marl/mfac.rst
+++ b/docs/source/documents/api/agents/marl/mfac.rst
@@ -42,8 +42,8 @@ MFAC
     :type obs: np.ndarray
     :param actions_mean: The mean values of actions.
     :type actions_mean: Tensor
-    :return: xxxxxx.
-    :rtype: xxxxxx
+    :return: hidden states and critic values.
+    :rtype: tuple
 
 .. py:function::
     xuance.torch.agent.mutli_agent_rl.mfac_agents.MFAC_Agents.train(i_step, kwargs)
@@ -98,8 +98,8 @@ MFAC
     :type obs: np.ndarray
     :param actions_mean: The mean values of actions.
     :type actions_mean: Tensor
-    :return: xxxxxx.
-    :rtype: xxxxxx
+    :return: hidden states and critic values.
+    :rtype: tuple
 
 .. py:function::
     xuance.tensorflow.agent.mutli_agent_rl.mfac_agents.MFAC_Agents.train(i_step, kwargs)
@@ -152,8 +152,8 @@ MFAC
     :type obs: np.ndarray
     :param actions_mean: The mean values of actions.
     :type actions_mean: Tensor
-    :return: xxxxxx.
-    :rtype: xxxxxx
+    :return: hidden states and critic values.
+    :rtype: tuple
 
 .. py:function::
     xuance.mindspore.agent.mutli_agent_rl.mfac_agents.MFAC_Agents.train(i_step, kwargs)
diff --git a/docs/source/documents/api/agents/marl/mfq.rst b/docs/source/documents/api/agents/marl/mfq.rst
index 334aa350a..999a82211 100644
--- a/docs/source/documents/api/agents/marl/mfq.rst
+++ b/docs/source/documents/api/agents/marl/mfq.rst
@@ -86,8 +86,8 @@ MFQ_Agents
 
     :param i_step: The i-th step during training.
     :type i_step: int
-    :param n_epoch: xxxxxx.
-    :type n_epoch: xxxxxx
+    :param n_epoch: Number of training epochs.
+    :type n_epoch: int
     :return: **info_train** - the information of the training process.
     :rtype: dict
 
@@ -130,8 +130,8 @@ MFQ_Agents
 
     :param i_step: The i-th step during training.
     :type i_step: int
-    :param n_epoch: xxxxxx.
-    :type n_epoch: xxxxxx
+    :param n_epoch: Number of training epochs.
+    :type n_epoch: int
     :return: **info_train** - the information of the training process.
     :rtype: dict
 
diff --git a/docs/source/documents/api/agents/marl/qmix.rst b/docs/source/documents/api/agents/marl/qmix.rst
index ede7e4725..e1956bf52 100644
--- a/docs/source/documents/api/agents/marl/qmix.rst
+++ b/docs/source/documents/api/agents/marl/qmix.rst
@@ -82,8 +82,8 @@ QMIX_Agents
 
     :param i_step: The i-th step during training.
     :type i_step: int
-    :param n_epoch: xxxxxx.
-    :type n_epoch: xxxxxx
+    :param n_epoch: Number of training epochs.
+    :type n_epoch: int
     :return: **info_train** - the information of the training process.
     :rtype: dict
 
@@ -118,12 +118,12 @@ QMIX_Agents
 .. py:function::
     xuance.mindspore.agents.mutli_agent_rl.qmix_agents.QMIX_Agents.train(i_step, n_epoch)
 
-    :param i_step: xxxxxx.
-    :type i_step: xxxxxx
-    :param n_epoch: xxxxxx.
-    :type n_epoch: xxxxxx
-    :return: xxxxxx.
-    :rtype: xxxxxx
+    :param i_step: The i-th step during training.
+    :type i_step: int
+    :param n_epoch: Number of training epochs.
+    :type n_epoch: int
+    :return: **info_train** - the information of the training process.
+    :rtype: dict
 
     
 .. raw:: html