Soil erosion is a pivotal environmental issue especially in monsoon-dominated regions, where intense rainfall events accelerate sediment transport and threaten agricultural productivity and water resources. This study applied the Revised Universal Soil Loss Equation (RUSLE) integrated with geospatial methods to assess rainfall-induced erosion in a sub-basin of the Upper Krishna River, India. The analysis focused on pre-flood (2018), flood (2019), and post-flood (2020) years to evaluate the influence of extreme rainfall on erosion dynamics. The model utilized rainfall data from the India Meteorological Department, soil attributes from the FAO database, SRTM-derived topography, and Landsat-based Normalized Difference Vegetation Index (NDVI) for vegetation cover. All RUSLE factors were computed in a GIS environment, and soil erosion maps were generated through pixel-wise multiplication. Results showed substantial interannual variability, mean soil loss was 4.2 t ha−1 yr−1 in 2018, rising sharply to 7.02 t ha−1 yr−1 in 2019, and decreasing to 5.6 t ha−1 yr−1 in 2020. Corresponding total erosion volumes were estimated at ~ 45, ~ 74.8, and ~ 60 million tons, respectively. Hotspot analysis indicated that central and western basin regions experienced severe erosion in 2019, consistent with the high rainfall erosivity factor 277 MJ mm ha−1 h−1 yr−1, compared with 122 and 244 MJ mm ha−1 h−1 yr−1 in 2018 and 2020. The findings identified rainfall as the key driver of erosion in the basin and demonstrate RUSLE’s reliability in mapping erosion risk. To mitigate future soil loss, conservation interventions such as afforestation, contour-based farming, vegetative buffers, and check dam construction are recommended, particularly in erosion hotspots identified by the study.