Abstract. With the raise of environmental protection awareness, applying models to control NPS (non-point source) pollution has become a key approach for environmental protection and pollution prevention and control in China. In this study, we implanted the semi-conceptual model SWAT (Soil and Water Assessment Tool) using information on rainfall runoff, land use, soil and slope. The model was used to quantify the spatial loading of NPS nutrient total nitrogen (TN) and total phosphorus (TP) to the Huntai River Watershed (HTRW) under two scenarios: without and with projected buffer zones of approximately 1 km within both banks of the Huntai, Taizi and Daliao river trunk streams and 5 km around the reservoirs. Current land-use types within the buffer zone were varied to indicate the natural ecology and environment. The Nash–Sutcliffe efficiency coefficient (E_NS) and R² for flow and predicted nutrient concentrations (TN and TP) in a typical hydrological station were both greater than 0.6, and the relative deviation (|Dv|) was less than 20 %. Under the status quo scenario (SQS), the simulated soil erosion in the HTRW per year was 811 kg/ha, and the output loadings of TN and TP were 19 and 7 kg/ha, respectively. The maximum loadings for TN and TP were 365 and 260 kg/ha, respectively. Under environmental protection scenarios (EPS), the TN and TP pollutant loadings per unit area were reduced by 26 % and 14 % annually, respectively. Loading analysis showed that land-use type is a key factor controlling NPS pollution. The NPS pollutant loading decreased under the simulated EPS, indicating that environmental protection measures may reduce the NPS pollutant loading in HTRW. The 22 % pollutant reduction under the EPS. We finally quantified the ratio of the land area lost to agricultural production compared with that lost to ecosystem services. We calculated the agricultural yield elasticity and concluded that the corresponding crop yield would be reduced by 2 % when the land area for ecosystem services in the basin increased by 1 % under the EPS.