diff --git a/paddle/fluid/imperative/README.md b/paddle/fluid/imperative/README.md
index 294c64b36ab3b7cb8a92fc8a60846fbc3ad1ef86..89543da752d7eebeafaf831b24166b65a7357738 100644
--- a/paddle/fluid/imperative/README.md
+++ b/paddle/fluid/imperative/README.md
@@ -21,8 +21,7 @@ class Layer(object):
     # ...
     return self.apply(inputs):
 
-
-  def apply(inputs):
+  def forward(inputs):
     # forward logic with paddle operators. backward auto-generated.
 
 
@@ -35,7 +34,8 @@ class PyLayer(core.PyLayer):
   def forward(inputs):
     # any forward logic implemented with numpy io.
 
-  @static method
+  @staticmethod
+  def backward(inputs):
     # any backward logic implemented with numpy io.
 ```
 
@@ -67,7 +67,6 @@ class Tracer {
 Lots of research already.
 https://autodiff-workshop.github.io/
 
-
 ## Tests
 
 * All op tests run once in static graph, once in imperative mode.
@@ -131,6 +130,7 @@ class MLP(fluid.imperative.Layer):
      out._backward()
 ```
 
+
 # Plan
 
 2.1，3 fulltime, Can run a few simple models. (Currently, 2 20% engs)
@@ -143,6 +143,7 @@ class MLP(fluid.imperative.Layer):
 
 12.1, 5 fulltime, Can compile to static graph, support more optimizations.
 
+
 # Discussion
 
 TODO.