Accurate landslide susceptibility evaluation is of great significance for disaster prevention and mitigation. Taking Loyang County as the study area, based on the Certainty Factor model (CF) susceptibility zoning, non-landslide points were selected after excluding the very high and high susceptibility zones, and the CF value was extracted as the input value of the support vector machine (SVM) model, and the optimal parameters were obtained by the Gray Wolf optimization algorithm to establish the CF-SVM model for predicting the study area, and also compared with the SVM model of randomly selected non-landslide points. The results show that the CF-SVM model covers 74.2% of the historical landslide points in the very high and high susceptibility areas, and the AUC reaches 0.95, both of which are higher than the SVM model, thus indicating that the CF-SVM model has a higher accuracy rate and proving the feasibility of selecting non-landslide units based on the CF model. It can provide a scientific basis for risk management in this region.