Abstract: In order to identify the main sources of water pollution in Zhangjiakou section of the Yongding River, and to determine the key control industries and priority control units of the basin, the receptor model was combined with the control unit division to establish the analysis method of pollution sources in the basin. Based on the water system distribution, administrative divisions and digital elevation model (DEM) data, the study area was divided into 16 control units. Cluster analysis was used to analyze the spatial characteristics of the water quality and pollution sources. Factor analysis and absolute principal component scores-multiple linear regression (APCS-MLR) receptor model were used to identify the source of pollution and calculate the contribution rate. The results showed that, according to the degree of water pollution, the study area could be classified into area A (middle and lower reaches of the Yang River and the Qingshui River) with heavy pollution, and area B (upper reaches of the Yang River and the Qingshui River and the entire Sanggan River) with light pollution. Area A was mainly affected by the mixture of industrial point sources and non-point sources. The contribution rate of industrial point sources was 43%, and that of agricultural plant pollution was 44%. Area B was mainly polluted by non-point sources. The main pollution was sewage from rural life and tourism (30%), agricultural planting (18%), and livestock breeding (17%). According to the spatial characteristics of pollution sources, the key control industries in area A were metallurgical and food manufacturing, and that in area B were mining and food manufacturing. Besides, the control units No.2 and 3 covering Yangyuan County, No.5 covering northern Zhuolu County and northern Yu County, and No.14 covering Wanquan County were identified as priority prevention and control units of source pollution. The research showed that the method combined source analysis with control unit division can reflect the spatial differentiation characteristics of the water quality and improve the source analysis ability.