Sponge city is currently a new urban rain flood management concept. It refers to the natural disaster that adapts to environmental changes and the natural disasters that is brought by rainwater. It is also known as the sponge, and the alias can also be called as “water elastic city.” The purpose of this article is, based on artificial intelligence, to study the analysis of wisdom management models in sponges. This paper first introduces the artificial intelligence algorithm and sponge city, then compares the traditional sponge city and the wisdom sponge city, then creates a LSTM neural network model, introducing artificial intelligence into sponge city intelligent dynamics in the analysis, and finally compares the rainfall data analysis to the ground. The experimental results show that, at different time points, the training results of rainfall data have shown significant regularity. The entire rainfall process exhibits a rise-decline-rise-increase trend of repetition. The maximum rainfall appeared at the 18th hour, and 7 obvious peaks occurred in 7, 11, 14, 16, 18, 21, and 23 hours.