diff --git a/python/paddle/fluid/layers/nn.py b/python/paddle/fluid/layers/nn.py
index e0588530e68252532684992c5e73054e38ac2fa4..5fead860a6fe55735055868008527e8fd775cc28 100644
--- a/python/paddle/fluid/layers/nn.py
+++ b/python/paddle/fluid/layers/nn.py
@@ -73,7 +73,6 @@ __all__ = [
     'autoincreased_step_counter',
     'unsqueeze',
     'lod_reset',
-    'relu',
     'clip',
     'clip_by_norm',
     'mul',
@@ -1826,51 +1825,6 @@ def lod_reset(x, y=None, target_lod=None):
     return out
 
 
-@deprecated(since="2.0.0", update_to="paddle.nn.functional.relu")
-def relu(x, name=None):
-    """
-    ${comment}
-
-    Args:
-        x(Variable): ${x_comment}
-        name(str, optional): The default value is None. Normally there is no
-            need for user to set this property. For more information, please
-            refer to :ref:`api_guide_Name`.
-
-    Returns:
-        Variable: ${out_comment}
-
-    Examples:
-
-        .. code-block:: python
-
-            import paddle.fluid as fluid
-            import numpy as np
-            in1 = np.array([[-1,0],[1,2.6]])
-            with fluid.dygraph.guard():
-                x1 = fluid.dygraph.to_variable(in1)
-                out1 = fluid.layers.relu(x1)
-                print(out1.numpy())
-                # [[0.  0. ]
-                #  [1.  2.6]]"""
-
-    if in_dygraph_mode():
-        return _C_ops.relu(x)
-    if _in_legacy_dygraph():
-        return _legacy_C_ops.relu(x)
-
-    check_variable_and_dtype(x, 'x', ['float16', 'float32', 'float64'], 'relu')
-
-    inputs = {'X': [x]}
-    helper = LayerHelper('relu', **locals())
-    dtype = helper.input_dtype(input_param_name='x')
-    out = helper.create_variable_for_type_inference(dtype)
-    helper.append_op(
-        type="relu", inputs={"X": helper.input('x')}, outputs={"Out": out}
-    )
-    return out
-
-
 def _logical_op(op_name, x, y, out=None, name=None, binary_op=True):
     if _non_static_mode():
         op = getattr(_legacy_C_ops, op_name)
diff --git a/python/paddle/fluid/reader.py b/python/paddle/fluid/reader.py
index 66427426cbbfff9821935e639ebaa129cf9a59b9..104ac8d864fb3bedd77d8f3c8d0a3adc421c4aca 100644
--- a/python/paddle/fluid/reader.py
+++ b/python/paddle/fluid/reader.py
@@ -1810,7 +1810,7 @@ class PyReader(DataLoaderBase):
                    paddle.batch(user_defined_reader, batch_size=BATCH_SIZE),
                    place)
                for image, label in py_reader():
-                   relu = fluid.layers.relu(image)
+                   relu = paddle.nn.functional.relu(image)
     """
 
     def __init__(
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/ifelse_simple_func.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/ifelse_simple_func.py
index f0ca20d3df735190c7b4b37c81daf53205d5e8e0..5d08ee96566e198830d5e2981c33b467562e24d6 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/ifelse_simple_func.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/ifelse_simple_func.py
@@ -62,7 +62,7 @@ def dyfunc_with_if_else2(x, col=100):
         # col = -1
         col = fluid.layers.fill_constant(shape=[1], value=-1, dtype="int64")
     if paddle.mean(x).numpy()[0] > x.numpy()[row][col]:
-        y = fluid.layers.relu(x)
+        y = paddle.nn.functional.relu(x)
     else:
         x_pow = paddle.pow(x, 2)
         y = paddle.tanh(x_pow)
@@ -163,7 +163,7 @@ def nested_if_else(x_v):
         w = fluid.layers.fill_constant([feat_size], dtype='float32', value=10)
         if y.numpy()[0] < 10:
             tmp = y * w
-            y = fluid.layers.relu(tmp)
+            y = paddle.nn.functional.relu(tmp)
             if paddle.mean(y).numpy()[0] < batch_size:
                 y = paddle.abs(y)
             else:
@@ -273,7 +273,7 @@ class NetWithControlFlowIf(fluid.dygraph.Layer):
                 # Create new var, but is not used.
                 x = 10
                 tmp = y * self.constant_vars['w']
-                y = fluid.layers.relu(tmp)
+                y = paddle.nn.functional.relu(tmp)
                 # Nested `if/else`
                 if y.numpy()[-1] < self.alpha:
                     # Modify variable of class
diff --git a/python/paddle/fluid/tests/unittests/ir/test_ir_fusion_group_pass.py b/python/paddle/fluid/tests/unittests/ir/test_ir_fusion_group_pass.py
index 47b65f5626ff6a6a53755701527fc948612db8f2..9b3fd4644cc51c251f07d1cae7a822dc0ba8a522 100644
--- a/python/paddle/fluid/tests/unittests/ir/test_ir_fusion_group_pass.py
+++ b/python/paddle/fluid/tests/unittests/ir/test_ir_fusion_group_pass.py
@@ -35,7 +35,7 @@ class FusionGroupPassTest(PassTest):
             tmp_0 = self.feed_vars[0] * self.feed_vars[1]
             tmp_1 = layers.mul(tmp_0, self.feed_vars[2])
             # subgraph with 2 op nodes
-            tmp_2 = layers.relu(tmp_0 + tmp_1)
+            tmp_2 = paddle.nn.functional.relu(tmp_0 + tmp_1)
 
         self.append_gradients(tmp_2)
 
@@ -146,7 +146,7 @@ class FusionGroupPassTestCastAndFP16(FusionGroupPassTest):
             tmp_3 = layers.mul(tmp_0, self.feed_vars[2])
             # subgraph with 4 op nodes
             tmp_3 = layers.cast(tmp_2, dtype="float16")
-            tmp_4 = layers.relu(tmp_1 + tmp_3)
+            tmp_4 = paddle.nn.functional.relu(tmp_1 + tmp_3)
             tmp_5 = layers.cast(tmp_4, dtype=dtype)
             tmp_3 = layers.cast(tmp_2, dtype=dtype)