Based on the observational data of precipitation at the stations of Zhejiang, relative bias(RB), root mean square error(RMSE), and correlation coefficient (CC) were used to evaluate the quasi-real-time product IMERG_ER V06B of GPM satellite during the landing periods of typhoon Chan-hom, Yagi, Lekima, Mitag, and Hagupit during 2015 to 2020. POD, FAR and CSI were used to evaluate the ability of precipitation retrieval of IMERG at different precipitation levels. The results show that the IMERG precipitation data of GPM can reflect the spatial distribution characteristics of precipitation, but its ability to capture the range of heavy precipitation is still insufficient. Precipitation during typhoon Yagi, Lekima, Mitag, and Hagupit were overestimated, while precipitation during typhoon Chan-hom were underestimated. IMERG has the ability to capture the change of precipitation of typhoon Hagupit, followed by typhoon Chan-hom, Yagi, and Lekima, and typhoon Mitag is the worst. According to the results of classification test, POD and CSI decrease, while FAR increases with the increase of precipitation, which indicates that IMERG has the best ability to identify light rain, but poor ability to capture heavy rain. Meanwhile, IMERG overestimated the possibility of heavy precipitation. Objectively, precipitation product of IMERG_ER V06B is insufficient to capture the precipitation intensity of typhoon, and the quantitative error of individual typhoon is large. Overall, it has a fair ability to capture the number and variation of peak and valley of typhoon rainfall, which can be applied to estimate the change of typhoon rainfall in complex terrain, due to its similar if not better ability of retrieving precipitation in mountainous area as in plain area.