This paper evaluates the prediction performance of the European Central Numerical Model (EC), the East China Regional Central Regional Numerical Model (WARMS) and the Jiangsu Regional Model (PWAFS) for 11 typical rainstorm processes in the Meiyu season in the Yangtze Huaihe River region in 2020 from three perspectives of precipitation deviation characteristics, prediction skills and object diagnosis analysis, and the advantages and disadvantages of each model were analyzed. Results show that: (1) the daily average accumulated precipitation observed in 24 hours is mainly distributed in the Dabie Mountains-Huaibei, Jiangsu Province and the Dabie Mountains-southern Anhui Province. The corresponding area with the maximum value of the model precipitation deviation has a good correspondence with the location of the main rain belt. Among them, EC has a significant dry deviation in the Dabie Mountains and southern Anhui Province, while in Huaibei and Jiangsu Province, there is a systematic north deviation. Both WARMS and PWAFS regional models have a large range of wet deviations in the upper reaches of the Dabie Mountains and southern Anhui mountains and in the lower reaches of Zhejiang Province, while they have a dry deviation in the Huaibei region of Jiangsu Province; (2) the scoring results of 24-hour prediction skills show that the TS score of EC for rainstorm level and below is the highest, while PWAFS is the best for heavy rainstorm level because of the high frequency of missing report of EC for heavy rainstorm. Comparing the two regional models of WARMS and PWAFS, it can be seen that PWAFS has lower missing report and false report rates at almost all levels than WARMS, so its TS score is also higher than that of WARMS; (3) through the diagnosis and analysis of MODE objects, it is found that EC has the most stable precipitation location prediction, PWAFS has the best prediction effect of precipitation intensity and range, but the forecast of rain band location is not stable enough. In general, the prediction performance of PWAFS is slightly better than that of WARMS. Compared with EC, PWAFS also has advantages in describing precipitation intensity and rain band range, but the prediction stability needs to be improved.