Abstract: Precipitation products with high spatial and temporal resolution and high accuracy are crucial for monitoring extreme precipitation events as well as preventing and mitigating disasters. Gauge station observations provide accurate point-scale precipitation but are insufficient for finely capturing spatial information on heavy precipitation induced by severe convection. Radar scanning can provide accurate precipitation information with high spatial and temporal resolution, but the accuracy of radar QPE (quantitative precipitation estimation) is vulnerable to various factors, such as observation accuracy and Z–R (radar reflectivity factor Z and rainfall rate R) relationship. Therefore, a radar–gauge merging algorithm is proposed in this paper to combine the advantages of gauge station observations and radar QPE. The algorithm includes three steps: Kriging interpolations of precipitation, LGC (local gauge-corrected) radar QPE, and radar–gauge merging QPE. First, the precipitation interpolation fields are obtained by the Kriging method based on the regional station observations. Subsequently, based on the LGC method, the accuracy of the radar QPE is improved by making systematic corrections. Finally, combined with the precipitation type, the radar–gauge merging QPE with high spatial and temporal resolution and high accuracy is produced by the radar–gauge merging algorithm. Three extreme precipitation events, namely, the 21·7 extreme precipitation in Zhengzhou, typhoon In-Fa, and the extreme precipitation in Suizhou in August 2021, are used to evaluate the performance of the radar–gauge merging algorithm. Results show that the new radar–gauge merging QPE outperforms the radar QPE product in terms of accuracy and characterizes the precipitation structure more finely than the Kriging interpolations of the gauge station observations in terms of the spatial distribution and time periods of the different extreme precipitation events. These results indicate the high accuracy and stability of the new radar–gauge merging algorithm and its ability to capture the distribution of extreme precipitation events.