Benggang (a fragmented erosion landform) have emerged as crucial factors restraining regional economic development, social stability, and ecological conservation due to their suddenness, explosiveness, and unpredictability. Monitoring the dynamic changes in distribution density, degree of development, and influencing factors of Benggang in large-scale natural environments can constitute preliminary references for studying the mechanisms, patterns, and influencing factors of Benggang in the natural environment, as well as providing insights for the coordination of regional socio-economic development and ecological environment governance. In this study, Sentinel-1/2 satellite constellation remote sensing images and other auxiliary data from 2016 to 2022 were combined with machine learning methods to identify Benggang erosion and measure the influencing factors of Benggang erosion in the middle and upper reaches of the Han River. The results indicate that: 1) The overall accuracy of the classification model in this study is 84%, with a Kappa coefficient of 0.8. The user's and producer's accuracy of Benggang erosion identification both exceed 95%, and their FScore is 0.97. However, smaller-scale Benggang erosion (< 100 m2) is easily confused with grassland and sparse vegetation, thereby reducing the Benggang identification accuracy. 2) Currently, the total area of Benggang erosion in the middle and upper reaches of the Han River is 435.5 km2. There are significant differences in Benggang erosion areas across counties, cities, and districts, with varying annual trends. In terms of spatial distribution, the Benggang erosion areas in Wuhua County, Xingning City, Dapu County, Fengshun County, Meixian District, Pingyuan County, Jiaoling County, and Meijiang District are 199.2, 70.1, 57.7, 50.1, 35.4, 14.9, 3.9, and 1.6 km2, respectively. In terms of time change trends, the average annual rates of change of Benggang erosion areas in Wuhua County, Xingning City, Dapu County, Fengshun County, Meixian District, Pingyuan County, Jiaoling County, and Meijiang District are 16.29, 6.51, 5.31, 4.94, 3.56, 1.39, 0.42, and 0.18 km2/a, respectively. 3) The occurrence probability of Benggang in the study area is significantly correlated (P < 0.001) with elevation, slope, vegetation coverage, geological type, population density, atmospheric pressure, rainfall, meridional wind speed, zonal wind speed, and wind speed. Within a certain range of variation in the influencing factors, elevation, slope, geological type, atmospheric pressure, meridional wind speed, zonal wind speed, and wind speed have a positive impact on Benggang occurrence probability in the study area, while vegetation coverage, population density, and rainfall have a negative impact on the occurrence of Benggang erosion. Furthermore, geological type, slope, atmospheric pressure, elevation, vegetation cover, population density and wind speed are the main influencing factors for the occurrence of Benggang in the study area. Therefore, strengthening the monitoring of these seven variables is necessary to enhance early warning capabilities for geological disasters due to Benggang in the study area.