diff --git a/mindspore/nn/layer/basic.py b/mindspore/nn/layer/basic.py
index b5bc2cbb99..68b90ed77b 100644
--- a/mindspore/nn/layer/basic.py
+++ b/mindspore/nn/layer/basic.py
@@ -188,10 +188,10 @@ class Dense(Cell):
         ValueError: If weight_init or bias_init shape is incorrect.
 
     Inputs:
-        - **input** (Tensor) - Tensor of shape :math:`(N, in\_channels)`.
+        - **input** (Tensor) - Tensor of shape :math:`(*, in\_channels)`.
 
     Outputs:
-        Tensor of shape :math:`(N, out\_channels)`.
+        Tensor of shape :math:`(*, out\_channels)`.
 
     Examples:
         >>> input = Tensor(np.random.randint(0, 255, [2, 3]), mindspore.float32)
@@ -200,7 +200,7 @@ class Dense(Cell):
         [[ 2.5246444   2.2738023   0.5711005  -3.9399147 ]
          [ 1.0739875   4.0155234   0.94188046 -5.459526  ]]
     """
-    @cell_attr_register(attrs=['has_bias', 'activation'])
+    @cell_attr_register(attrs=['has_bias', 'activation', 'in_channels', 'out_channels'])
     def __init__(self,
                  in_channels,
                  out_channels,
diff --git a/tests/st/ops/ascend/test_dense.py b/tests/st/ops/ascend/test_dense.py
index 65b1eb400f..acb90b362e 100644
--- a/tests/st/ops/ascend/test_dense.py
+++ b/tests/st/ops/ascend/test_dense.py
@@ -31,6 +31,18 @@ class Net(nn.Cell):
     def construct(self, x):
         return self.dense(x)
 
+class MultiLayerDense(nn.Cell):
+    def __init__(self):
+        super(MultiLayerDense, self).__init__()
+        self.dense1 = nn.Dense(in_channels=256, out_channels=512)
+        self.dense1 = nn.Dense(in_channels=512, out_channels=1024)
+
+    @ms_function
+    def construct(self, x):
+        x = self.dense1(x)
+        x = self.dense2(x)
+        return x
+
 
 def test_net():
     x = np.random.randn(32, 2048).astype(np.float32)
@@ -46,3 +58,11 @@ def test_net_ND():
     output = net(Tensor(x))
     print(x)
     print(output.asnumpy())
+
+
+def test_net_multilayer():
+    x = np.random.randn(16, 32, 256).astype(np.float32)
+    net = MultiLayerDense()
+    output = net(Tensor(x))
+    print(x)
+    print(output.asnumpy())
diff --git a/tests/st/ops/gpu/test_dense.py b/tests/st/ops/gpu/test_dense.py
new file mode 100644
index 0000000000..0b69869266
--- /dev/null
+++ b/tests/st/ops/gpu/test_dense.py
@@ -0,0 +1,65 @@
+# Copyright 2019 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+import numpy as np
+
+import mindspore.context as context
+import mindspore.nn as nn
+from mindspore import Tensor
+
+context.set_context(device_target="GPU")
+
+
+class Net(nn.Cell):
+    def __init__(self):
+        super(Net, self).__init__()
+        self.dense = nn.Dense(2048, 1001)
+
+    def construct(self, x):
+        return self.dense(x)
+
+class MultiLayerDense(nn.Cell):
+    def __init__(self):
+        super(MultiLayerDense, self).__init__()
+        self.dense1 = nn.Dense(in_channels=256, out_channels=512)
+        self.dense1 = nn.Dense(in_channels=512, out_channels=1024)
+
+    def construct(self, x):
+        x = self.dense1(x)
+        x = self.dense2(x)
+        return x
+
+
+def test_net():
+    x = np.random.randn(32, 2048).astype(np.float32)
+    net = Net()
+    output = net(Tensor(x))
+    print(x)
+    print(output.asnumpy())
+
+
+def test_net_ND():
+    x = np.random.randn(2, 332, 2048).astype(np.float32)
+    net = Net()
+    output = net(Tensor(x))
+    print(x)
+    print(output.asnumpy())
+
+
+def test_net_multilayer():
+    x = np.random.randn(16, 32, 256).astype(np.float32)
+    net = MultiLayerDense()
+    output = net(Tensor(x))
+    print(x)
+    print(output.asnumpy())