Natural Zoning for Highways is the important basis of highway network planning and establishment of related specifications and standards. Determination method for influence coefficients of comprehensive zoning indices analysis model is the key to increase quantization degree of comprehensive zoning indices analysis. Based on the theories of fuzzy clustering and rough set, this paper put forward the method to reasonably determine the influence coefficients of single elements and calibrate the analysis model of comprehensive zoning indices through analyzing the calculation elements data by random sampling method. Highway surface configuration index(HSCI) was analyzed with this method. The map of HSCI reflects the influence degree of surface configuration on highway engineering in different regions in China correctly. Analysis result shows that the method improves the quantization degree of comprehensive zoning indices analysis.