Abstract: Near-real-time satellite precipitation retrievals have the advantages of wide coverage, spatial continuity, short latency and open access, serving as an important precipitation data source that is globally available. For the 20 July 2021 extreme rainfall event in Henan Province, the performance of four GPM near-real-time satellite precipitation products—IMERG early, IMERG late, GSMaP NOW and GSMaP Gauge NOW—is comprehensively evaluated in this study. Based on the observations of 116 ground meteorological gauges and the interpolated rainfall fields, their skills to characterize heavy rainfall are compared.①The results show that the underestimation of accumulated rainfall by IMERG early and IMERG late is about 20%, while GSMaP NOW and GSMaP Gauge NOW overestimates it by about 35% and 70% respectively. However, the latter two are found to be easier to detect accumulated rainfall of above 500mm. ②In terms of rainfall process, the four GPM datasets are able to detect hourly precipitation events, but fail to capture major rainfall peaks. The estimation error of hourly rain rate is prominent, and it is negatively correlated with the ground observed rainfall magnitude. Specifically, the four products tend to overestimate hourly rainfall event that is less than 10 mm/h and underestimate rainfall event exceeding 30 mm/h. ③As for spatial pattern, the evaluation metrics of all the datasets show strong temporal fluctuations. The spatial correlation coefficient and volume critical success index of IMERG data are generally better than those of GSMaP data, but the latter is more sensitive to high-magnitude precipitation events. ④For IMERG products, it is found that the accuracy of IMERG late is greatly improved upon IMERG early; for GSMaP products, GSMaP Gauge NOW has a better detection skill of high-intensity rainfall compared against GSMaP NOW, but it also shows an increased estimation error of low-magnitude rainfall. This study has deepened the understanding of the performance of various GPM near-real-time satellite precipitation products, and provided critical feedbacks for improving the GPM-era satellite precipitation retrieval algorithms and enhancing their skills to monitor extreme precipitation.