Abstract: Hongze Lake is the fourth largest freshwater lake in China,and also an important water passage in the east route of“South-toNorth Water Diversion Project”. Its water environment quality is very important for the water supply guarantee of its surrounding and north water-receiving areas. In this paper,a wavelet analysis based long short-term memory neural network water quality prediction model（WALSTM）is established. Firstly,the water quality data of Huaihe entering Hongze Lake is preprocessed,and then it is decomposed into highfrequency and low-frequency signals and input into LSTM model for single factor water quality prediction. Then,the water quality change of Hongze Lake is comprehensively predicted and evaluated by single factor water quality prediction result driven fuzzy neural network based on T-S. The results show that wavelet analysis can capture the characteristics of water quality data better,and the combination of wavelet analysis and LSTM water quality prediction model can greatly improve the accuracy of water quality prediction and that the comprehensive application of WA-LSTM model and fuzzy neural network can effectively solve the problem that single factor prediction can not reflect the overall situation of water quality. The research methods and conclusions of this paper can provide technical support and practical reference for water quality monitoring and management and water resources regulation in Hongze Lake.