The construction of sponge city is of great significance. How to better predict urban drainage becomes a key point. Aiming at the volatility and nonlinear characteristics of urban drainage, a city water quantity prediction algorithm based on fuzzy neural network prediction is proposed. Firstly, the gray model is used to construct the urban drainage prediction model, and the objective prediction regression function is established based on the concept of sponge city. Secondly, the sliding window is constructed by setting breakpoints, and the data of the gray model is segmented in real time to obtain real-time statistical features of the data. We construct a function between the sequence error prediction and the compression ratio, and use the error prediction sequence to determine the segmentation point to improve the robustness of the detection process. Finally, through the urban drainage prediction simulation experiment, it is shown that the proposed algorithm can effectively improve the prediction accuracy and prediction efficiency of urban drainage, and better reflect the overall trend.