diff --git a/docs/api/paddle/vision/Overview_cn.rst b/docs/api/paddle/vision/Overview_cn.rst
index d339fbe8952..fef448aa247 100644
--- a/docs/api/paddle/vision/Overview_cn.rst
+++ b/docs/api/paddle/vision/Overview_cn.rst
@@ -54,6 +54,12 @@ paddle.vision 目录是飞桨在视觉领域的高层API。具体如下：
     " :ref:`vgg13 <cn_api_paddle_vision_models_vgg13>` ", "13层的VGG模型"
     " :ref:`vgg16 <cn_api_paddle_vision_models_vgg16>` ", "16层的VGG模型"
     " :ref:`vgg19 <cn_api_paddle_vision_models_vgg19>` ", "19层的VGG模型"
+    " :ref:`DenseNet <cn_api_paddle_vision_models_DenseNet>` ", "DenseNet模型"
+    " :ref:`densenet121 <cn_api_paddle_vision_models_densenet121>` ", "121层的DenseNet模型"
+    " :ref:`densenet161 <cn_api_paddle_vision_models_densenet161>` ", "161层的DenseNet模型"
+    " :ref:`densenet169 <cn_api_paddle_vision_models_densenet169>` ", "169层的DenseNet模型"
+    " :ref:`densenet201 <cn_api_paddle_vision_models_densenet201>` ", "201层的DenseNet模型"
+    " :ref:`densenet264 <cn_api_paddle_vision_models_densenet264>` ", "264层的DenseNet模型"
     " :ref:`InceptionV3 <cn_api_paddle_vision_models_InceptionV3>` ", "InceptionV3模型"
     " :ref:`inception_v3 <cn_api_paddle_vision_models_inception_v3>` ", "InceptionV3模型"
 
diff --git a/docs/api/paddle/vision/models/DenseNet_cn.rst b/docs/api/paddle/vision/models/DenseNet_cn.rst
new file mode 100644
index 00000000000..c9a433277cc
--- /dev/null
+++ b/docs/api/paddle/vision/models/DenseNet_cn.rst
@@ -0,0 +1,35 @@
+.. _cn_api_paddle_vision_models_DenseNet:
+
+DenseNet
+-------------------------------
+
+.. py:class:: paddle.vision.models.DenseNet(layers=121, bn_size=4, dropout=0., num_classes=1000)
+
+ DenseNet模型，来自论文 `"Densely Connected Convolutional Networks" <https://arxiv.org/abs/1608.06993>`_ 。
+
+参数
+:::::::::
+  - **layers** (int, 可选) - densenet的层数。默认值：121。
+  - **bn_size** (int，可选) - 中间层growth rate的拓展倍数。默认值：4。
+  - **dropout** (float, 可选) - dropout rate。默认值：0.。
+  - **num_classes** (int，可选) - 类别数目，即最后一个全连接层输出的维度。默认值：1000。
+  - **with_pool** (bool，可选) - 是否定义最后一个全连接层之前的池化层。默认值：True。
+
+返回
+:::::::::
+DenseNet模型，Layer的实例。
+
+代码示例
+:::::::::
+.. code-block:: python
+
+    import paddle
+    from paddle.vision.models import DenseNet
+
+    # build model
+    densenet = DenseNet()
+
+    x = paddle.rand([1, 3, 224, 224])
+    out = densenet(x)
+
+    print(out.shape)
diff --git a/docs/api/paddle/vision/models/desnet121_cn.rst b/docs/api/paddle/vision/models/desnet121_cn.rst
new file mode 100644
index 00000000000..9d4369b0978
--- /dev/null
+++ b/docs/api/paddle/vision/models/desnet121_cn.rst
@@ -0,0 +1,34 @@
+.. _cn_api_paddle_vision_models_densenet121:
+
+densenet121
+-------------------------------
+
+.. py:function:: paddle.vision.models.densenet121(pretrained=False, **kwargs)
+
+ 121层的densenet模型，来自论文 `"Densely Connected Convolutional Networks" <https://arxiv.org/abs/1608.06993>`_ 。
+
+参数
+:::::::::
+  - **pretrained** (bool，可选) - 是否加载在imagenet数据集上的预训练权重。默认值：False。
+
+返回
+:::::::::
+densenet121模型，Layer的实例。
+
+代码示例
+:::::::::
+.. code-block:: python
+
+    import paddle
+    from paddle.vision.models import densenet121
+
+    # build model
+    model = densenet121()
+
+    # build model and load imagenet pretrained weight
+    # model = densenet121(pretrained=True)
+
+    x = paddle.rand([1, 3, 224, 224])
+    out = model(x)
+
+    print(out.shape)
diff --git a/docs/api/paddle/vision/models/desnet161_cn.rst b/docs/api/paddle/vision/models/desnet161_cn.rst
new file mode 100644
index 00000000000..e52d9267fbb
--- /dev/null
+++ b/docs/api/paddle/vision/models/desnet161_cn.rst
@@ -0,0 +1,34 @@
+.. _cn_api_paddle_vision_models_densenet161:
+
+densenet161
+-------------------------------
+
+.. py:function:: paddle.vision.models.densenet161(pretrained=False, **kwargs)
+
+ 161层的densenet模型，来自论文 `"Densely Connected Convolutional Networks" <https://arxiv.org/abs/1608.06993>`_ 。
+
+参数
+:::::::::
+  - **pretrained** (bool，可选) - 是否加载在imagenet数据集上的预训练权重。默认值：False。
+
+返回
+:::::::::
+densenet161模型，Layer的实例。
+
+代码示例
+:::::::::
+.. code-block:: python
+
+    import paddle
+    from paddle.vision.models import densenet161
+
+    # build model
+    model = densenet161()
+
+    # build model and load imagenet pretrained weight
+    # model = densenet161(pretrained=True)
+
+    x = paddle.rand([1, 3, 224, 224])
+    out = model(x)
+
+    print(out.shape)
diff --git a/docs/api/paddle/vision/models/desnet169_cn.rst b/docs/api/paddle/vision/models/desnet169_cn.rst
new file mode 100644
index 00000000000..44ccbeaa690
--- /dev/null
+++ b/docs/api/paddle/vision/models/desnet169_cn.rst
@@ -0,0 +1,34 @@
+.. _cn_api_paddle_vision_models_densenet169:
+
+densenet169
+-------------------------------
+
+.. py:function:: paddle.vision.models.densenet169(pretrained=False, **kwargs)
+
+ 169层的densenet模型，来自论文 `"Densely Connected Convolutional Networks" <https://arxiv.org/abs/1608.06993>`_ 。
+
+参数
+:::::::::
+  - **pretrained** (bool，可选) - 是否加载在imagenet数据集上的预训练权重。默认值：False。
+
+返回
+:::::::::
+densenet169模型，Layer的实例。
+
+代码示例
+:::::::::
+.. code-block:: python
+
+    import paddle
+    from paddle.vision.models import densenet169
+
+    # build model
+    model = densenet169()
+
+    # build model and load imagenet pretrained weight
+    # model = densenet169(pretrained=True)
+
+    x = paddle.rand([1, 3, 224, 224])
+    out = model(x)
+
+    print(out.shape)
diff --git a/docs/api/paddle/vision/models/desnet201_cn.rst b/docs/api/paddle/vision/models/desnet201_cn.rst
new file mode 100644
index 00000000000..0144e8ba045
--- /dev/null
+++ b/docs/api/paddle/vision/models/desnet201_cn.rst
@@ -0,0 +1,34 @@
+.. _cn_api_paddle_vision_models_densenet201:
+
+densenet201
+-------------------------------
+
+.. py:function:: paddle.vision.models.densenet201(pretrained=False, **kwargs)
+
+ 201层的densenet模型，来自论文 `"Densely Connected Convolutional Networks" <https://arxiv.org/abs/1608.06993>`_ 。
+
+参数
+:::::::::
+  - **pretrained** (bool，可选) - 是否加载在imagenet数据集上的预训练权重。默认值：False。
+
+返回
+:::::::::
+densenet201模型，Layer的实例。
+
+代码示例
+:::::::::
+.. code-block:: python
+
+    import paddle
+    from paddle.vision.models import densenet201
+
+    # build model
+    model = densenet201()
+
+    # build model and load imagenet pretrained weight
+    # model = densenet201(pretrained=True)
+
+    x = paddle.rand([1, 3, 224, 224])
+    out = model(x)
+
+    print(out.shape)
diff --git a/docs/api/paddle/vision/models/desnet264_cn.rst b/docs/api/paddle/vision/models/desnet264_cn.rst
new file mode 100644
index 00000000000..ddc20b2fac5
--- /dev/null
+++ b/docs/api/paddle/vision/models/desnet264_cn.rst
@@ -0,0 +1,34 @@
+.. _cn_api_paddle_vision_models_resnet264:
+
+densenet264
+-------------------------------
+
+.. py:function:: paddle.vision.models.densenet264(pretrained=False, **kwargs)
+
+ 264层的densenet模型，来自论文 `"Densely Connected Convolutional Networks" <https://arxiv.org/abs/1608.06993>`_ 。
+
+参数
+:::::::::
+  - **pretrained** (bool，可选) - 是否加载在imagenet数据集上的预训练权重。默认值：False。
+
+返回
+:::::::::
+densenet264模型，Layer的实例。
+
+代码示例
+:::::::::
+.. code-block:: python
+
+    import paddle
+    from paddle.vision.models import densenet264
+
+    # build model
+    model = densenet264()
+
+    # build model and load imagenet pretrained weight
+    # model = densenet264(pretrained=True)
+
+    x = paddle.rand([1, 3, 224, 224])
+    out = model(x)
+
+    print(out.shape)