Engineering information is a very precious information resources, and it is of great significance to predict the engineering cost. In this paper, using BP neural network algorithm through to complex, decentralized completed engineering data collecting and analyzing statistics and draw all kinds of traffic engineering of the quantity of consumption, and the market price of the overall trend. This can provide the basis for the investment estimation, calculation of engineering cost, engineering quotation and contract adjustment, and provide the main basis for the new engineering project decision, construction and design.