diff --git a/paddle/fluid/API.spec b/paddle/fluid/API.spec
index 11ccef30f607061705284247b59acea930a4615e..3f9cc30a18a5851ddf8052a5b379db945e17b728 100644
--- a/paddle/fluid/API.spec
+++ b/paddle/fluid/API.spec
@@ -109,7 +109,7 @@ paddle.fluid.layers.reduce_prod (ArgSpec(args=['input', 'dim', 'keep_dim', 'name
 paddle.fluid.layers.sequence_first_step (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', '2b290d3d77882bfe9bb8d331cac8cdd3'))
 paddle.fluid.layers.sequence_last_step (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', 'c16a892f44f7fe71bfa5afc32d3f34ce'))
 paddle.fluid.layers.sequence_slice (ArgSpec(args=['input', 'offset', 'length', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', 'fdcea0e8b5bc7d8d4b1b072c521014e6'))
-paddle.fluid.layers.dropout (ArgSpec(args=['x', 'dropout_prob', 'is_test', 'seed', 'name', 'dropout_implementation'], varargs=None, keywords=None, defaults=(False, None, None, 'downgrade_in_infer')), ('document', 'dc7042734c6d8b8ce97321f017f01d6f'))
+paddle.fluid.layers.dropout (ArgSpec(args=['x', 'dropout_prob', 'is_test', 'seed', 'name', 'dropout_implementation'], varargs=None, keywords=None, defaults=(False, None, None, 'downgrade_in_infer')), ('document', 'f1dd22f7351f7f9853212958e0d8aa7a'))
 paddle.fluid.layers.split (ArgSpec(args=['input', 'num_or_sections', 'dim', 'name'], varargs=None, keywords=None, defaults=(-1, None)), ('document', '652625345c2acb900029c78cc75f8aa6'))
 paddle.fluid.layers.ctc_greedy_decoder (ArgSpec(args=['input', 'blank', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', 'ebbf2adbd79683dc93db03454dfa18c2'))
 paddle.fluid.layers.edit_distance (ArgSpec(args=['input', 'label', 'normalized', 'ignored_tokens'], varargs=None, keywords=None, defaults=(True, None)), ('document', '97f0262f97602644c83142789d784571'))
diff --git a/paddle/fluid/operators/dropout_op.cc b/paddle/fluid/operators/dropout_op.cc
index 7c64eadffffbafe0d7f92ec345d75a92c4fef58e..65c2ff6415c1d51fdc05d6014da589678761b676 100644
--- a/paddle/fluid/operators/dropout_op.cc
+++ b/paddle/fluid/operators/dropout_op.cc
@@ -71,7 +71,7 @@ class DropoutOpMaker : public framework::OpProtoAndCheckerMaker {
         "1. downgrade_in_infer(default), downgrade the outcome at inference "
         "time"
         "   train: out = input * mask"
-        "   inference: out = input * dropout_prob"
+        "   inference: out = input * (1.0 - dropout_prob)"
         "2. upscale_in_train, upscale the outcome at training time, do nothing "
         "in inference"
         "   train: out = input * mask / ( 1.0 - dropout_prob )"
diff --git a/python/paddle/fluid/layers/nn.py b/python/paddle/fluid/layers/nn.py
index dbe495b75c8876de3c3f0fb0abe0357089413254..28655314d30fef5c35bd9ea1898e41c0231d7417 100644
--- a/python/paddle/fluid/layers/nn.py
+++ b/python/paddle/fluid/layers/nn.py
@@ -1348,7 +1348,7 @@ def dropout(x,
                                         1. downgrade_in_infer(default), downgrade the outcome at inference
 
                                            - train: out = input * mask
-                                           - inference: out = input * dropout_prob
+                                           - inference: out = input * (1.0 - dropout_prob)
 
                                            (mask is a tensor same shape with input, value is 0 or 1
                                            ratio of 0 is dropout_prob)