With the rapid growth of the demand for college and university funding, the financial management of colleges and universities from the traditional meaning of the risk-free state to the risky mode of change, for the college and university financial risk of accurate and reasonable prevention and control has become an important issue that needs to be resolved at this stage. This paper selects a university finance as the research object, designs 12 financial indicators as sample data, processes the financial risk indicator system through principal component analysis, and obtains 8 principal factor components as the input data of the risk prediction model. Then the particle swarm algorithm is combined with BP neural network to overcome the defects of BP neural network as a way to strengthen its prediction accuracy of financial risk. Through simulation experiments, comparative analysis of prediction rates of different models, it is found that the PSO-BP prediction model achieves an identification accuracy of 91.7% for 60 test samples, which improves the identification correctness by 23.7% and 8.4% compared with the traditional BP model and GA-BP model, respectively. It confirms that the PSO-BP neural network model has a higher prediction rate and is effective in introducing university finance for risk prediction. Finally, the article proposes an optimization strategy for the path of university finance information technology construction, in order to improve the effectiveness of university finance information technology construction.