This study aims to assess the spatial patterns of soil erosion and sediment dynamics in the Shalmanrud watershed, eastern Gilan Province, Iran, using an integrated modeling framework combining the RUSLE and WaTEM/SEDEM models within the Google Earth Engine (GEE) platform. Rainfall erosivity (R) was estimated from 20 years of monthly precipitation records using the Modified Fournier Index, while soil erodibility (K) was derived from soil texture and organic matter content based on the Wischmeier equation. The topographic factor (LS) was calculated from a 30-m resolution digital elevation model using the Desmet and Govers algorithm, and the cover-management factor (C) was obtained from NDVI values derived from Landsat 8 imagery. Owing to the absence of effective soil conservation measures, the support practice factor (P) was assumed to be unity across the watershed. The spatially distributed RUSLE outputs were subsequently used as inputs to the WaTEM/SEDEM model to simulate sediment production, transport, and deposition processes. Results indicate that annual soil erosion rates range from near zero to approximately 44.15 t ha⁻¹ yr⁻¹, with nearly 72% of the watershed classified as very low erosion (<5 t ha⁻¹ yr⁻¹). WaTEM/SEDEM simulations estimate total annual sediment production at about 73,829 t yr⁻¹ (≈1.46 t ha⁻¹ yr⁻¹), whereas internal sediment deposition reaches approximately 603,840 t yr⁻¹. These findings suggest that the Shalmanrud watershed predominantly functions as a sediment sink rather than a sediment-exporting system.