In order to effectively avoid existed in the development of scientific research project, technical risk, market risk, financial risk, legal risk, organizing risk, and a series of wind, ensure the smooth progress of the research and development of scientific research project reduce project failure rate, improve the project achievement conversion rate, artificial intelligence technology has been widely used in the existing environment, artificial intelligence, neural network technology is introduced into the software development has become a possibility. In this paper, the idea of introducing BP neural network technology into the development process of sharing platform for scientific research projects is proposed. That the platform can be a lot of characteristic parameters of all kinds of scientific research project, utilizing the nonlinear mapping ability of BP neural network, adaptive ability and generalization ability of discrete data generation model, the implementation of a new project from multidimensional generate reliable outlook, to strengthen the scientific research project sharing platform project in the future the reliability of the prediction. It has been proved that the prediction of the prospect of scientific research projects based on BP neural network technology is more accurate and efficient than the prediction by the relevant personnel of scientific research projects, and the practical value of the sharing platform of scientific research projects is also improved.