This study evaluates the reliability of satellite-derived datasets for estimating hydrological soil erosion estimation by validating Google Earth Engine (GEE)–based RUSLE outputs against field-measured and conventionally GIS-processed data in the Nekor watershed (northern Morocco) over a 40-year period (1983–2023). The watershed, which supplies the Mohamed Ben Abdelkrim El Khattabi Dam, has experienced accelerated sedimentation due to intensified surface runoff, marl–schist degradation, and declining vegetation cover. Soil erosion was estimated using the RUSLE model within two parallel frameworks: (i) a ground-based GIS approach incorporating locally calibrated field data, and (ii) a multi-source satellite-driven workflow implemented in GEE. Long-term temporal trends were analyzed to assess erosion dynamics and evaluate the effectiveness of soil conservation measures introduced in the early 2000s.Results from the field-based GIS approach indicate a 16.5% reduction in mean annual soil loss, decreasing from 47.87 to 39.96 t ha⁻¹ yr⁻¹ over the study period. Comparable trends were observed in the GEE-derived estimates, which declined from 46.08 to 37.04 t ha⁻¹ yr⁻¹. Statistical validation demonstrates very strong agreement between the two approaches, with coefficients of determination (R²) of 0.99 and 0.98 for the two analyzed periods. Error metrics further confirm the robustness of the satellite-based methodology, with high Nash–Sutcliffe efficiency values (NSE = 0.97 in 2003 and 0.96 in 2023), low root mean square error (RMSE = 12.43 and 18.22 t ha⁻¹ yr⁻¹), and low mean absolute error (MAE = 10.00 and 15.61 t ha⁻¹ yr⁻¹), accompanied by minimal bias (+2.62 and +3.01), indicating only slight overestimation by GEE.Spatial analysis further reveals high concordance in the identification of critical erosion hotspots, particularly in steep upstream areas characterized by fragile lithology and sparse vegetation cover. The novelty of this study lies in the long-term (four-decade) quantitative validation of GEE-based RUSLE outputs against locally calibrated field and GIS datasets in a Mediterranean mountain watershed, providing rare empirical evidence of the accuracy, consistency, and scalability of cloud-based erosion modeling.Overall, the results demonstrate that satellite-derived datasets processed within GEE offer a robust, scalable, and cost-effective alternative for long-term soil erosion monitoring and watershed management, particularly in data-scarce Mediterranean environments.