Abstract: The methods of inventory analysis, equivalent load method, cluster analysis and geographic information system (GIS) technology were used to analyze the pollution sources, pollutants and their spatial distribution characteristics of agricultural non-point sources (NPS) in Beijing in 2016. The results showed that the main sources of agricultural NPS pollution in Beijing were mainly rural residents and livestock and poultry. It was suggested that the pollution control of rural residents and livestock and poultry in the future agricultural NPS pollution control should be strengthened. Agricultural NPS pollution showed strong regional differentiation characteristics, with aquaculture pollution being mainly concentrated in Pinggu District, and rural living pollution being the most serious in Fangshan District. TN, TP and COD pollution indexes in Fangshan, Tongzhou, Shunyi, Changping and Daxing Districts all exceeded the average level of the whole city, and Fangshan District was the highest. According to the spatial clustering analysis of agricultural NPS pollution risk, the whole Beijing could be divided into three primary zones: high, medium and light risk zones. In addition, based on the main types of pollution sources in each zone, it could be divided into 8 secondary sub-zones. Therefore, the agricultural NPS pollution could be classified and controlled, and the corresponding prevention and control countermeasures were proposed.