Abstract:In order to improve the accuracy of the basin’s runoff prediction and increase the efficiency of operation of hydropower stations, the control catchment of the Xiaowan Hydropower Station was selected as a study area. Based on landuse data from the years 1986, 1995, and 2000 taken from remote sensing satellites and multi-year data of reservoir inflow of the Xiaowan Hydropower Station, with the runoff coefficient used as the dependent variable, and arable land, forest land, building land, and other underlying surface factors as independent variables, the regression relation between the underlying surface factors and the runoff coefficient was analyzed with the SPSS regression model and a non-linear regression model. With consideration of the influence of the rainfall in the previous year on the runoff in the following year, the method of rainfall-runoff simulation was adopted. The three methods were used to simulate the runoff coefficient. The results show that the non-linear regression model fits the runoff coefficient better than the other two models, and it can be applied to the prediction of runoff and enhancement of the utilization of hydropower in medium and small-sized basins.