Water resource prediction is the basic guarantee of urban safe water use, while campus water use prediction is the basis of urban water use planning and management. Aiming at solving the uncertainty of campus water consumption caused by several factors, a prediction model of campus water consumption based on grey genetic BP neural network is proposed. The model carries out grey relational analysis on the data of campus water consumption, and adds genetic algorithm to optimize BP neural network. After residual calculation, the predicted value of the output interval is obtained. This model can fully extract information from small samples and add genetic optimization neural network to solve the problem that neural network cannot automatically optimize. The results of Matlab simulation of campus water use interval data show that the predicted data and actual data are basically consistent with a simulation accuracy 90.32%, and it verified the feasibility of the method, the prediction method has some reference significance.