Abstract. NPS (Non-point source) pollution has become a key impact element to watershed environment at present. With the development of technology, application of models to control NPS pollution has become a very common practice for resource management and Pollutant reduction control in the watershed scale of China. The SWAT (Soil and Water Assessment Tool) model is a semi-conceptual model, which was put forward to estimate pollutant production & the influences on water quantity-quality under different land development patterns in complex watersheds. Based on the overview of published papers with application of SWAT, the study topics is mainly focus on nutrients, sediments, impoundment & wetlands, hydrologic characteristics, climate change impact, and land-use change impacts. SWAT model was constructed based on rainfall runoff and land use type. The migration-transformation processes of agricultural NPS pollution was simulated and calculated based on the SWAT model. Besides, the loadings and distribution traits of NPS pollutants were also systematically analyzed based on the model. The model was used to quantify the spatial loading intensities of NPS nutrient TN (Total Nitrogen) and TP (Total Phosphorus) to HTRW (Huntai River Watershed) under two scenarios (without & with buffer zones). The SWAT model was validated using actual monitoring information as well as the physical properties of the underlying substrate, hydrology, meteorology and pollutant sources in the HTRW. Scenario settings are mainly based on the changes of surface runoff and sediments, climate and land-use change from different spatial scales, and climatic/physiographic zones. About 1 km within both banks of the trunk streams of the Huntai, Taizi and Daliao rivers, and 5 km surrounding the reservoirs were defined as buffer zones. Existing land use type within the buffer zone was changed to reflect the natural environment. The output of pollutant production under the EPS (Environmental Protection Scenarios) was calculated based on the status quo scenario. Under the status quo scenario, the annual mean modulus of soil erosion in the HTRW was 811 kg/ha, and the output intensities of TN & TP were 19 & 7 kg/ha, respectively. For the unit area, the maximal loading intensities for TN & TP were 365.36 & 259.83 kg/ha, respectively. In terms of spatial distribution, TN & TP loading varied substantially. Under the EPS, the magnitude of N & P production from arable land decreased to a certain degree, and the TN & TP pollution loading per unit area were reduced by 5 & 1 kg/ha annually, respectively. In comparison, the quantity of NPS pollutant production under the EPS was reduced by 21.9 % compared with the status quo scenario, and the quantities of TP & TN decreased by 10.4 % & 25.9 %, respectively. These changes suggested a clear reduction in the export loading of agricultural NPS pollution. Loading intensities analysis showed that land use type is one key factor for controlling NPS pollution. The NPS pollutant loading decreased under the EPS, which showed that environmental protection measure could effectively cut down NPS pollutant loading in HTRW. SWAT was used to assess the reduction of agricultural NPS pollutant. However, SWAT model requires a large amount of data about the watershed being modeled; the data inaccuracy and local factors would impact the accuracy of the SWAT model. To determine the pollutant reduction under different land development patterns, and examine uncertainty of sensitivity parameters, SWAT model in China has wide range of potential application.