Abstract: As the successor of TMPA,the Global Precipitation Mission（GPM）IMERG（Integrated Multi-satellitE Retrievals for GPM）products provide a source of precipitation data with greater coverage,longer time series,and higher accuracy and spatial and temporal resolution. Based on gauging daily rainfall from 824 meteorological stations,this paper adopts five extreme precipitation indices（R95 p,R95 pTOT,Rx1 day,Rx5 day and R25 mm）and three statistical metrics（CC,Bias and rRMSE）to evaluate the extreme precipitation detection capability of IMERG Final product（hereafter IMERG）and compare with the results of TMPA 3 B42 V7 product（hereafter TMPA 3 B42）.The result shows that:（1） The accuracy of IMERG generally outperforms TMPA 3 B42 product with higher CC values and lower rRMSE values of each extreme rainfall indices.（2） Except for the underestimation of IMERG in R95 p,both IMERG and TMPA 3 B42 estimates show overestimation on all five extreme rainfall indices.（3） In extreme rainfall detection,the accuracy of the two satellite products has decreased under complex topography conditions,but IMERG exhibits smaller range and lower level of bias than that of TMPA 3 B42.（4） For analyzing the frequency distribution of precipitation with different intensities,the distribution of IMERG is more consistent with gauge precipitation compared to that of TMPA 3 B42,but there is underestimation at higher rainfall intensity（＞150 mm/d）in both IMERG and TMPA 3 B42.