The present invention relates to a method for determining a surface runoff yield in a vegetation-covered area. The present invention improves and integrates a water conservation model with a Zhang's model based on remote sensing data. The present invention constructs a new method for calculating a surface runoff yield in a vegetation-covered area on a spatial pixel scale based on a water balance equation of the vegetation-covered area. This method utilizes real-time dynamic multi-temporal remote sensing data to calculate a vegetation canopy interception water storage, a vegetation litterfall interception water storage, a soil water storage change, a vegetation water conservation, a vegetation evapotranspiration and a vegetation runoff yield. The method realizes the long-term dynamic estimation of the surface runoff yield in regional and global vegetation-covered areas on a spatial pixel scale. It has the advantages of being efficient, fast, accurate, and applicable to large-scale vegetation-covered areas.