diff --git a/python/paddle/fluid/layers/nn.py b/python/paddle/fluid/layers/nn.py
index 53d35f92e11c2bea12e87b8434fa8e63921f8a9f..d8c1432de611674d729f1dfe069bcd478cbf9ec9 100755
--- a/python/paddle/fluid/layers/nn.py
+++ b/python/paddle/fluid/layers/nn.py
@@ -9520,36 +9520,35 @@ def pow(x, factor=1.0, name=None):
 @templatedoc()
 def stanh(x, scale_a=0.67, scale_b=1.7159, name=None):
     """
+    stanh activation.
 
-    ${comment}
-    Args:
-        x(${x_type}): ${x_comment}
-        scale_a(${scale_a_type}|2.0 / 3.0): ${scale_a_comment}
-        scale_b(${scale_b_type}|1.7159): ${scale_b_comment}
-        name(str|None): A name for this layer(optional). If set None, the layer
-                        will be named automatically.
+    .. math::
+
+        out = b * \\frac{e^{a * x} - e^{-a * x}}{e^{a * x} + e^{-a * x}}
+
+    Parameters:
+        x (Tensor): The input Tensor with data type float32, float64.
+        scale_a (float, optional): The scale factor a of the input. Default is 0.67.
+        scale_b (float, optional): The scale factor b of the output. Default is 1.7159.
+        name (str, optional): Name for the operation (optional, default is None).
+            For more information, please refer to :ref:`api_guide_Name`.
 
     Returns:
-        output(Tensor): ${out_comment}.
+        A Tensor with the same data type and shape as ``x`` .
 
     Examples:
-
         .. code-block:: python
 
             import paddle
 
-            data = paddle.rand(shape=[3, 3], dtype='float32')
-            output = paddle.stanh(data, scale_a=0.67, scale_b=1.72)
-            print(data)
-            # [[0.19412413, 0.66871136, 0.77059180],
-            #  [0.89738929, 0.35827777, 0.60592669],
-            #  [0.66346580, 0.78424633, 0.46533889]]
-            print(output)
-            # [[0.22245567, 0.72288811, 0.81671900],
-            #  [0.92525512, 0.40512756, 0.66227961],
-            #  [0.71790355, 0.82885355, 0.51953089]]
+            x = paddle.to_tensor([1.0, 2.0, 3.0, 4.0])
+            out = paddle.stanh(x, scale_a=0.67, scale_b=1.72) # [1.00616539, 1.49927628, 1.65933108, 1.70390463]
 
     """
+
+    if in_dygraph_mode():
+        return core.ops.stanh(x, 'scale_a', scale_a, 'scale_b', scale_b)
+
     check_variable_and_dtype(x, 'x', ['float16', 'float32', 'float64'], 'stanh')
 
     helper = LayerHelper('stanh', **locals())
diff --git a/python/paddle/fluid/tests/unittests/test_activation_op.py b/python/paddle/fluid/tests/unittests/test_activation_op.py
index e969184628949ab92ce5cf6ca0233cfc6c8edd1b..f0bb15ae93bb2e6e008f8d47987e1027c9285511 100755
--- a/python/paddle/fluid/tests/unittests/test_activation_op.py
+++ b/python/paddle/fluid/tests/unittests/test_activation_op.py
@@ -1906,18 +1906,30 @@ class TestPow_factor_tensor(TestActivation):
         self.assertRaises(TypeError, fluid.layers.pow, x=in4, factor=factor_1)
 
 
+def ref_stanh(x, scale_a=0.67, scale_b=1.7159):
+    out = scale_b * np.tanh(x * scale_a)
+    return out
+
+
 class TestSTanh(TestActivation):
+    def get_scale_a(self):
+        return 0.67
+
+    def get_scale_b(self):
+        return 1.7159
+
     def setUp(self):
         self.op_type = "stanh"
         self.init_dtype()
+        scale_a = self.get_scale_a()
+        scale_b = self.get_scale_b()
 
         np.random.seed(1024)
         x = np.random.uniform(0.1, 1, [11, 17]).astype(self.dtype)
-        scale_a = 2.0 / 3.0
-        scale_b = 1.7159
-        out = scale_b * np.tanh(x * scale_a)
+        # The same reason with TestAbs
+        out = ref_stanh(x, scale_a, scale_b)
 
-        self.inputs = {'X': OpTest.np_dtype_to_fluid_dtype(x)}
+        self.inputs = {'X': x}
         self.attrs = {'scale_a': scale_a, 'scale_b': scale_b}
         self.outputs = {'Out': out}
 
@@ -1927,17 +1939,85 @@ class TestSTanh(TestActivation):
         self.check_grad(['X'], 'Out')
 
 
-class TestSTanhOpError(unittest.TestCase):
+class TestSTanhScaleA(TestSTanh):
+    def get_scale_a(self):
+        return 2.0
+
+
+class TestSTanhScaleB(TestSTanh):
+    def get_scale_b(self):
+        return 0.5
+
+
+class TestSTanhAPI(unittest.TestCase):
+    # test paddle.nn.stanh
+    def get_scale_a(self):
+        return 0.67
+
+    def get_scale_b(self):
+        return 1.7159
+
+    def setUp(self):
+        np.random.seed(1024)
+        self.x_np = np.random.uniform(-1, 1, [10, 12]).astype('float32')
+        self.scale_a = self.get_scale_a()
+        self.scale_b = self.get_scale_b()
+        self.place=paddle.CUDAPlace(0) if core.is_compiled_with_cuda() \
+            else paddle.CPUPlace()
+
+    def test_static_api(self):
+        paddle.enable_static()
+        with paddle.static.program_guard(paddle.static.Program()):
+            x = paddle.fluid.data('X', [10, 12])
+            out = paddle.stanh(x, self.scale_a, self.scale_b)
+            exe = paddle.static.Executor(self.place)
+            res = exe.run(feed={'X': self.x_np}, fetch_list=[out])
+        out_ref = ref_stanh(self.x_np, self.scale_a, self.scale_b)
+        for r in res:
+            self.assertEqual(np.allclose(out_ref, r), True)
+
+    def test_dygraph_api(self):
+        paddle.disable_static(self.place)
+        x = paddle.to_tensor(self.x_np)
+        out = paddle.stanh(x, self.scale_a, self.scale_b)
+        out_ref = ref_stanh(self.x_np, self.scale_a, self.scale_b)
+        for r in [out]:
+            self.assertEqual(np.allclose(out_ref, r.numpy()), True)
+        paddle.enable_static()
+
+    def test_fluid_api(self):
+        paddle.enable_static()
+        with fluid.program_guard(fluid.Program()):
+            x = fluid.data('X', [10, 12])
+            out = fluid.layers.stanh(x, self.scale_a, self.scale_b)
+            exe = fluid.Executor(self.place)
+            res = exe.run(feed={'X': self.x_np}, fetch_list=[out])
+        out_ref = ref_stanh(self.x_np, self.scale_a, self.scale_b)
+        self.assertEqual(np.allclose(out_ref, res[0]), True)
+
     def test_errors(self):
-        with program_guard(Program()):
+        paddle.enable_static()
+        with paddle.static.program_guard(paddle.static.Program()):
             # The input type must be Variable.
-            self.assertRaises(TypeError, fluid.layers.stanh, 1)
+            self.assertRaises(TypeError, paddle.stanh, 1)
             # The input dtype must be float16, float32, float64.
-            x_int32 = fluid.data(name='x_int32', shape=[12, 10], dtype='int32')
-            self.assertRaises(TypeError, fluid.layers.stanh, x_int32)
+            x_int32 = paddle.fluid.data(
+                name='x_int32', shape=[12, 10], dtype='int32')
+            self.assertRaises(TypeError, paddle.stanh, x_int32)
             # support the input dtype is float16
-            x_fp16 = fluid.data(name='x_fp16', shape=[12, 10], dtype='float16')
-            fluid.layers.stanh(x_fp16)
+            x_fp16 = paddle.fluid.data(
+                name='x_fp16', shape=[12, 10], dtype='float16')
+            paddle.stanh(x_fp16)
+
+
+class TestSTanhAPIScaleA(TestSTanhAPI):
+    def get_scale_a(self):
+        return 2.0
+
+
+class TestSTanhAPIScaleB(TestSTanhAPI):
+    def get_scale_b(self):
+        return 0.5
 
 
 def ref_softplus(x, beta=1, threshold=20):
diff --git a/python/paddle/nn/layer/activation.py b/python/paddle/nn/layer/activation.py
index edab5660517e33a8ba553194c3653a98ec3547d0..1d1c7becea0f48a90e2075fb897298b0b9a38c59 100644
--- a/python/paddle/nn/layer/activation.py
+++ b/python/paddle/nn/layer/activation.py
@@ -536,7 +536,7 @@ class Sigmoid(layers.Layer):
 
     .. math::
 
-        Sigmoid(x) = \frac{1}{1 + e^{-x}}
+        Sigmoid(x) = \\frac{1}{1 + e^{-x}}
 
     Parameters:
         name (str, optional): Name for the operation (optional, default is None). For more information, please refer to :ref:`api_guide_Name`.
@@ -551,15 +551,11 @@ class Sigmoid(layers.Layer):
 
         .. code-block:: python
 
-          import numpy as np
           import paddle
 
-          paddle.disable_static()
-          input_data = np.array([1.0, 2.0, 3.0, 4.0]).astype('float32')
           m = paddle.nn.Sigmoid()
-          x = paddle.to_tensor(input_data)
-          output = m(x)
-          print(output.numpy()) # [0.7310586, 0.880797, 0.95257413, 0.98201376]
+          x = paddle.to_tensor([1.0, 2.0, 3.0, 4.0])
+          out = m(x) # [0.7310586, 0.880797, 0.95257413, 0.98201376]
     """
 
     def __init__(self, name=None):