Modelling sediment yield in large river basins using hydrological models is often limited by the scarcity of field-based input data. In particular, the vegetation and management factor of the Modified Universal Soil Loss Equation (MUSLE-C) is frequently oversimplified, leading to reduced accuracy in sediment simulations. This study applied remote sensing to map the spatial distribution of MUSLE-C and incorporated it into the SWAT model to assess sediment yield responses to climate variability in the Upper Tekeze Basin. Model performance was evaluated using the coefficient of determination (R²), percent bias (PBIAS), and Nash–Sutcliffe efficiency (NSE), while trends and associations between climate and sediment variables were analysed using the Pettitt–Kendall trend test and Pearson correlation analysis. Results demonstrate that LSUA-based MUSLE-C mapping enabled improved SWAT simulations, achieving NSE = 0.84, PBIAS = 10%, and R² = 0.79. Climate analysis revealed no statistically significant trends in rainfall, maximum temperature, or minimum temperature at seasonal or annual scales. In contrast, simulated sediment yields at the Tekeze Dam exhibited a significant increasing trend, highlighting the limited influence of broad-scale climate trends on sediment transport. These findings underscore the need for higher-resolution climate and sediment data to better understand and manage erosion processes in the basin.