Abstract: Background Traditional research on rainfall erosivity has largely depended on the data from ground‑based meteorological stations. However, the limited spatial density of these stations and discontinuous records hinder high-resolution and real-time monitoring of rainfall erosivity over large, topographically complex areas. Consequently, there is an urgent need to utilize satellite precipitation products, which offer superior spatiotemporal coverage and data continuity, to evaluate changes in rainfall erosivity. Methods Based on GPM IMERG and CMORPH satellite precipitation data, and using grid rainfall erosivity data from high-density stations as a benchmark, this study employed a grid-by-grid correction method to generate the rainfall erosivity data for the middle reaches of the Yellow River basin from 2001 to 2020. Then the differences in its temporal and spatial variation characteristics were analyzed. Results The results showed that the CMORPH satellite data estimated the rainfall erosivity significantly higher than GPM IMERG in the middle reaches of the Yellow River basin. The average rainfall erosivity value estimated by CMORPH over the study period was 1.3 times that of GPM IMERG. Spatially, high rainfall erosivity values from GPM IMERG were located in the Beiluo River and Jing River basins, while the high-value areas identified by CMORPH were more extensive. The grid correction coefficients indicated that GPM IMERG underestimated the rainfall erosivity in 57.4% of the area within the middle reaches of the Yellow River basin, and these underestimated areas were primarily concentrated in core sediment source areas in the middle reaches of the Yellow River basin. CMORPH, conversely, overestimated rainfall erosivity in 75.8% of the area. The derived grid correction coefficients can provide key support for real-time inversion and monitoring of rainfall erosivity using satellite precipitation data. Analysis of the satellite data after correction revealed a significant upward trend in rainfall erosivity in the middle reaches of the Yellow River from 2001 to 2020. The growth rates of rainfall erosivity based on GPM IMERG and CMORPH were 11.2 and 12.7 MJ·mm/(hm²·h·a²), respectively. The number of grids exhibiting significantly increased rainfall erosivity based on CMORPH was 1.6 times that of GPM IMERG. Rainfall erosivity varied significantly among sub-basins in the middle reaches of the Yellow River basin. The mean rainfall erosivity in the Qin River basin was the highest, and the rainfall erosivity in the Fen River basin and the mainstream from Hekou to Longmen increased significantly, with a significantly higher inter-annual variation rate than other sub-basins. Conclusions The research results provide a scientific basis for dynamic monitoring of soil and water loss and the optimization of prevention and control measures in the middle reaches of the Yellow River basin.