The spatial characteristic of non-point source pollution in the Liaohe River was studied. Coupling the remote sensing data and non-point source (NPS) models, a method of assessing NPS pollution by pixel unit was developed, aiming to analyse the NPS pollution characteristic of Liaohe River basin in 2010, in turn to identify the main polluted areas and prevention measures. The work will provide technical supports for pollution prevention in Liaohe River basin. The results showed that in 2010, the total discharge of total nitrogen (TN) was 1.03 x 10(5) t, the total phosphorus (TP) was 6.8 x 10(3) t, the chemical oxygen demand (COD) was 1.31 x 10(5) t and the ammonia nitrogen (NH+4 -N) was 1. 8 x 10(4) t. The main pollution source of NPS was from agriculture. The contributions of NPS pollution to water quality were 67.4% , 76.4% , 39.4% and 21.9% for TN, TP, COD and NH+4 -N, respectively. The south of Liaohe River basin was the most serious polluted area, followed by the northeast areas. In this research, a method was build to estimate the NPS loads based on remote sensing pixel and the spatial characteristic of non-point source pollution in Liaohe River in 2010 was analysed, which will provide support for pollution prevention in Liaohe River.