Abstract: Soil heavy metals come from a wide range of sources, whose pollution risks are influenced by varieties of factors. Usually, there is a nonlinear relationship between the pollution risk and the influencing factor. Logistic regression method could overcome the limitation of ordinary linear regression method, and has good application potential in the identification of influencing factors and the prediction of soil pollution risk. Taking the soil in a certain region strongly influenced by humans in Zhejiang Province as the object, the influencing factors of soil pollution risk and predicting the pollution risk were investigated based on the binary logistic regression model, with the independent variables of slope, aspect, altitude, population density, road distance, weighted distance of non-mining industry and distance to mining area, and the dependent variable of risk classification of soil Cu pollution. The results showed that the risk of soil Cu pollution was significantly affected by population density, distance to mining area, road distance, and weighted distance of non-mining industry, but there was no simple linear relationship between it and population density, distance to mining area and road distance. The additive effect of intensive mines dramatically changed the distance effect on the soil pollution risk. The high risk area of soil Cu pollution was mainly located in the Cu mining area and its influence area of the tailings pond, the medium risk area were located in the gold deposit of fracture zones, iron producing area and transportation influenced zone, and the low risk area was located in the areas of high and low density population.