Based on hourly observation from 74 meteorological stations in Zhejiang Province, the models of ordinary linear regression and those considering spatial heterogeneity are adopted to explore the spatial estimation and impact factors of summer afternoon precipitation in Zhejiang Province in this study. It is found that the summer afternoon precipitation in Zhejiang Province shows a decreasing trend from the southwest to northeast. The two main precipitation centers in the south and north form the southwest-northeast major rain band. Geographical, topographic and meteorological factors jointly affect the spatial distribution of summer afternoon precipitation in Zhejiang Province. The model of geographically neural network weighted regression (GNNWR) based on neural network and considering the spatial heterogeneity works significantly better than the models of ordinary linear regression and the geographically weighted regression, in terms of model performance and spatial distribution pattern of estimated precipitation, which suggests that the GNNWR model has the value of further application in the field of meteorology.