Based on the principle of BP neural network in and on the analysis of the characteristics ofhighway engineering,this paper identified seven engineering characteristics as the input vector of theneural network,build the highway engineering cost prediction model is based on BP neural network,combined with MATLAB neural network toolbox to design,and has the engineering as an example.Bytraining,correction,and to verify the model,it is proven that BP neural network can effectively improvethe accuracy of the prediction,has very good practical value.