The regulation of hydrological processes is greatly aided by vegetation, particularly in the areas of redistribution of precipitation, soil moisture movement, and modification of runoff generation and confluence conditions. The spatial arrangement of vegetation, in addition to its quantity, is essential for sustainability by improving hydrological mitigation. Numerous conventional rainfall-runoff models, such as the Soil Conservation Service (now Natural Resources Conservation Service) Curve Number (SCS-CN) method, one of the most widely used methods for predicting surface runoff, only took into account the type and coverage of the vegetation while ignoring the vegetation pattern factor. This will unavoidably lead to uncertainty in runoff predictions. In this study, the vegetation pattern indices of the mean flow path lengths index (MFLI), which was used to take into account the spatial position and distribution of vegetation at slope scale, was combined with the traditional SCS-CN method to derive a new CN value. Data from experimental plots in 48 sites of China were used to assess the proposed method's reliability. The optimized parameters were then applied to the remaining data from test plots in 5 typical sites. The proposed approach outperformed the original SCS-CN method, according to the results, increasing the model efficiencies to 85.88% and 83.51% for the calibration and validation scenarios, respectively. Therefore, an accurate runoff predicted for various vegetation pattern conditions at slope scale in China is suggested by the proposed method that takes into account the influence of the spatial position and distribution of vegetation.