Abstract: Abstract: This study aimed to simulate the processes of non-point source pollution transport and to identify the corresponding influencing factors using the SWAT model (soil and water assessment tool) in the Qingshitan Irrigation District (QID) of Lijiang River Basin. Field measurements were conducted in Huixian experiment area (HEA) of QID to monitor runoff and the concentration of total nitrogen (TN) and total phosphorus (TP) at the outlet of HEA. In the study measured monthly data were used to calibrate and validate the SWAT model, in which the thickness, bulk density, saturated hydraulic conductivity and available water content of soil were selected to quantify the karst development in HEA. Then scenario analysis was carried out to study the impacts of the irrigation and fertilization in paddy fields, the attributes (area and water storage capacity) of wetlands, and karst development degree on runoff, TN and TP emissions from HEA. The results showed that the SWAT model performed well in simulating runoff, TN and TP emission in the study area with all the relationship coefficient (R2) and Nash-Suttcliffe efficiency (NSE) between the simulated and measured data higher than 0.70 and 0.63, respectively. The scenario simulation showed that lower fertilization and irrigation could reduce TN and TP emissions by 2.46%-19.18% and 1.86%-14.21%, respectively. When fertilization was decreased by 30%, TN and TP emissions from HEA declined by 11.45% and 8.98%, respectively; and the reductions were 7.79% and 5.81%, respectively, when irrigation water was reduced by 30%, indicating that it was more efficient to decrease TN and TP emissions by reducing fertilization than irrigation and that TN emission was more prone to be reduced by optimized fertilization and irrigation than TP emission. However, the effects of reducing fertilization and irrigation in paddy fields on improving the water quality of outflow were relatively limited. It was needed to excavate the potential of pollutant purification of wetlands in HEA to meet the water quality standard of water function division in Guilin. The scenario simulation showed that variations of area and/or water storage capacity of wetland also affected TN and TP emissions from HEA. Runoff was hardly affected by changing area and water storage capacity of wetlands, however, 12.40% of TN and 10.44% of TP emission were decreased with the area and water storage capacity of wetlands rising by 50%, while TN and TP emission increased by 11.75% and 10.97%, respectively, with decreasing 50% of area and water storage capacity of wetlands. And the efficiency of reducing TN and TP emission of wetland area was higher than water storage capacity. Moreover, it had synergistic effects on TN and TP emissions by simultaneously increasing the area and water storage of wetlands compared with only changing one attribute. Soil parameters describing karst development degree had diverse influences on runoff, TN and TP emissions from subbasins 2, 4, 5. Thinner soil thickness and smaller bulk density decreased runoff, TN and TP emissions, while available water content and saturated hydraulic conductivity demonstrated the opposite effects. The conclusion indicated that more developed karst landscape exacerbated runoff, TN and TP emission. Besides, the runoff was more sensitive to soil thickness and available water content than bulk density and saturated hydraulic conductivity, while TN and TP emissions were more sensitive to soil thickness and bulk density. This study showed that controlling fertilization reasonably, preventing wetland shrinking and water capacity degrading, and maintaining and improving soil structure in the karst area was helping to alleviate nitrogen and phosphorus emissions in the karst irrigation area of Lijiang River Basin, which guided optimizing local agricultural measures and soil and water management.