Abstract: There exists the problems of single scenario, large amount of data, multiple links involved, complex models, cumbersome operations and unclear governance effects in the process of identifying the key source areas of agricultural non-point source pollution (ANSP). To solve these problems, an information system that could identify and simulate the treatment of the key source areas of ANSP was constructed by using GIS technology combined with InVEST water yield and nutrient transfer rate model. The key source areas identification function was designed under three scenarios, based on the river inflow load, potential runoff concentration and load to water yield ratio, respectively, and the distribution of key source areas under the three scenarios was simulated. The water yield, nutrient transfer rate and loss load sub-models of InVEST were integrated. A simulation function for the treatment of key source areas of ANSP was designed, which could visually display the treatment effects of the key source areas identified under the three scenarios. Taking Haihe River Basin as an example, the software was used to identify and simulate the treatment of the key source areas of non-point source of total nitrogen (TN) in farmland. The results showed that: 1) The key source areas identified based on the river inflow load scenario were relatively scattered, and distributed in most of the water systems except Yongding River and Ziya River. Based on the potential runoff concentration scenario, the distribution of key source areas was relatively concentrated, mostly distributed in the areas from the middle to the south, and the rest were distributed in the southeast. The distribution of key source areas identified based on the load-to-yield ratio scenario of the river basin was highly concentrated, ranging from central to southern of the area. 2) The scenarios based on potential runoff concentration and the ratio of load to water yield had similar effects in the simulation of TN and total phosphorus (TP) governance, and were significantly better than the scenario based on the inflow load. Combined with the above three scenarios set in the software, it could reduce the impact of complex geographic environment, improve the identification efficiency of key source areas of ANSP, and enhance the scientific nature of governance decision-making for the areas. The software is practical.