The objective of this study was to obtain the 24-h precipitation frequency estimates for return periods of 1-1000 years in Shaoguan, South China, to provide scientific suggestions for flood disaster warning and urban flood control planning. Annual Maximum Precipitation (AMP) respectively from observational rainfall and Integrated Multi-satellitE Retrievals for GPM (IMERG) gridded rainfall which can supplement precipitation information in areas without rain gauges was used for the calculation of estimates using Regional L-Moments Analysis (RLMA) method. Then the quantiles at grids were corrected by Mean Bias Correction (MBC) method with the estimates at sites. The results showed that there has more and stronger extreme precipitation in the northwest of Shaoguan. The number of homogeneous regions as well as the spatial distribution of parameters (Cv , Cs and Ck) for two datasets is different. When the return periods are larger than 100 years, the estimates at sites are also larger than that at grids and the difference will increase as return period increases. After modified, the quantiles from IMERG are more reasonable and combine the spatial distribution characteristic of estimates from two cases, which indicates that the rainstorm high-risk area of Shaoguan is in the southwest to the central, as well as in the west-central.