Abstract: Soil erosion is the result of coordinated interaction of multiple influencing factors.Accurately assessing the spatial distribution of regional soil erosion and its influencing factors is of great significance for in-depth exploration on the mechanisms of soil erosion and soil and water conservation efforts.Based on RUSLE model, multi-source remote sensing data, factors such as elevation, slope, land use type, vegetation coverage, and annual total rainfall were selected to systematically analyze the dynamic changes in soil erosion in the controlled watershed at the dam site of the Baihetan Hydropower Station from 2014 to 2021.The main driving mechanisms influencing soil erosion were also investigated.The results showed: ① The soil erosion modulus exhibited a continuous decline over the years (R²=0.82), and erosion intensity was mainly slight and mild levels, and showing significant spatial clustering (P < 0.01);② Land use type and slope were the main driving factors affecting soil erosion (q>18%), and their interactions exhibited nonlinear enhancement or two-factor enhancement; ③ Shrubland, grassland, and forestland were the primary sources of soil erosion, and areas of intense and moderate erosion were transfored to slight and mild erosion areas over time.