Abstract: This study focuses on landslide susceptibility assessments in Luoping County, where 9 evaluation factors, including engineering rock group, slope, slope aspect, elevation, undulation, curvature, landform type, distance from rivers, and distance from fault, were selected as the research variables. After conducting collinearity diagnosis and correlation analysis, the information value method was applied to calculate the information value for each classification level of the evaluation factors. Quantitative weights for each evaluation factor were determined using the AHP and logistic regression methods, leading to the construction and comparison of three susceptibility evaluation models: information value, and weighted information value, and information-logistic regression coupled model. The results were categorized into four grades -- none, low, medium, and high – using the GIS-based natural breakpoint method, and their accuracy was validated using ROC curves. The results show that the AUC values of the three evaluation models were 0.757, 0.723 and 0.852 respectively, with the information-logistic regression coupled model demonstrating the highest accuracy. Moreover, the model results were in good agreement with the distribution of landslide geological disaster points. The respective areas (classification ratios) for the none, low, medium, and high categories were 771.1 km² (25.55%), 836.6 km² (27.73%), 864.36 km² (28.64%), and 545.94 km² (18.08%).