Abstract: Currently, the best management practices for farmland are the primary means of controlling non-point source pollution in the Erhai Lake Basin. In this study, we aimed to assess the effectiveness of various management strategies in mitigating non-point source pollution in the basin and identify the best management practices for farmland. Using a rate-determined and validated Soil and Water Assessment Tool (SWAT) model, we simulated the effects of reducing nitrogen and phosphorus losses under 16 different scenarios, including reduction of nitrogen and phosphorus fertilizers, adjustment of tillage pattern, establishment of vegetated buffer zone, establishement of grass planted stream, and implementing comprehensive measures. The calibrated SWAT model demonstrated satisfactory performance, with a coefficient of determination (R²) of at least 0.67 (exceeding the threshold of 0.60) and a Nash-Sutcliffe model efficiency coefficient (E_ns) of at least 0.56 (surpassing the required E_ns of 0.50), confirming its suitability for simulating best management practices in the Erhai Lake Basin. Our analysis revealed that reduction of nitrogen and phosphorus fertilizers by 10%, 20%, and 50% could lead to nitrogen-loss reductions of 9.1%, 13.9% and 14.6%, respectively, and phosphorus-loss reductions of 1.7%, 2.5% and 5.4%, respectively. Among the various tillage practices examined, deep tillage was the most effective measure in controlling nitrogen and phosphorus losses, with reductions of 13.8% and 14.1%, respectively. The establishment of vegetated buffer zones resulted in nitrogen-loss reductions ranging from 9.3% to 24.8% and phosphorus-loss reductions ranging from 10.4% to 27.2% as the width of the vegetated buffer zone increased from 1 to 10 m. The implementation of 1−10 m long grass planted stream led to nitrogen-loss reductions of 21.3%–39.6% and phosphorus-loss reductions of 24.9%–45.2%. Compared with individual measures, comprehensive measures demonstrated substantially higher efficiency in reducing nitrogen and phosphorus losses. For example, the combination of a 20% reduction in fertilizer application with deep plowing achieved a nitrogen-loss reduction of 27.4%, which increased to 43.7% with the addition of a 5-m-wide vegetated buffer zone and 41.6% with the addition of a 20-m-long grass planted stream. The corresponding phosphorus-loss reduction efficiencies under these scenarios were 19.2%, 39.7%, and 61.1%. In conclusion, compared with individual strategies, comprehensive measures substantially reduced the non-piont source pollution loads in the watershed. Our results quantify the impacts of various farmland management practices on non-piont pollution in the Erhai Lake Basin and provide a theoretical foundation for effective prevention and control measures against non-piont pollution.