Accurately and effectively predicting precipitation is conducive to the planning of agricultural production development,water resources management and prevention of natural disasters,and is more significant in arid and semi-arid areas.This paper uses the precipitation data of Qingyang City from January 2023 to January 2024,iteratively removes unimportant features based on recursive feature elimination in the packaging method,and then uses the random forest model to analyze and predict the data.The results show that by integrating the two methods,the model can have good prediction performance,and can make a good prediction for precipitation time and precipitation amount in Qingyang City.The research content in this paper also has reference value for precipitation prediction in other cities,and is also of great significance to the rational use of local water resources and the promotion of sustainable local social and economic development.