<p>Soil erosion is known as one of the main factors contributing to soil health degradation and as one of the most common hazards associated with land degradation worldwide. Soil erosion threatens the sustainability of our world. The mechanisms, the process, and their spatial patterns of soil erosion can be analyzed and assessed by models. Models are efficient tools for understanding nature and the human impact. To model soil erosion, in this study, the SWAT and RUSLE models were evaluated to estimate soil loss and sediment yield in the Zayandeh-Rood dam watershed in central Iran. For runoff and sediment estimation using the SWAT model, the coefficient of determination (R²) and Nash-Sutcliffe efficiency (NSE) for runoff calibration and validation were obtained as 0.69, 0.64, 0.72, and 0.66, respectively. Similarly, these coefficients for sediment calibration and validation were found to be 0.62, 0.60, 0.58, and 0.55, respectively. The estimated sediment yield at the outlet point, based on the SWAT model’s calibration data, was 4.7 t ha^-1 yr^-1. By implementation of the slope-based sediment delivery ratio (SDR), average soil erosion at the watershed was estimated at 14.16 t ha^-1 yr^-1. Additionally, the prediction of eroded soil using the RUSLE model was estimated as about 177,908 t yr^-1 with an average value of 7.97 t ha^-1 yr^-1. The midterm data at the hydrometric station showed sediment production of 3.93 t ha^-1 yr^-1; with exerting SDR, the real soil erosion was estimated at approx. 11.83 t ha^-1 yr^-1. The results indicated that the RUSLE and SWAT models underestimated and overestimated, respectively, soil loss compared to measured data. The soil erosion classification of the study area was developed using the SWAT and RUSLE models, identifying regions with severe and very severe erosion classes. These areas require urgent biological and mechanical conservation measures to reduce sediment delivery to the Zayandeh-Rood Dam reservoir.</p>