Abstract: To address the issues of limited interpretability and excessive dependence on landslide data in statistical-based landslide susceptibility models,the Fuzzy Preference Ratio(FPR)has been introduced for landslide susceptibility assessment. This approach involves the dynamic updating of landslide susceptibility assessments as landslide data is updated,while environmental factors remain constant over a short period. The stability coefficient of susceptibility(w) is established to quantify the variation in the susceptibility class of slope units during this dynamic update. Three typical towns in the Wanzhou District of the Three Gorges Reservoir Area were chosen as the study area. A total of 113 landslides in the area were categorized into three groups in an 8 ︰ 1 ︰ 1 ratio to simulate the dynamic updating of landslide data. The information value model and the artificial neural network(ANN)were selected for comparison,with landslide susceptibility assessments conducted using data from the three groups to study the dynamic updating of landslide susceptibility evaluation. The findings indicate that after two updates of landslide data,more than 50% of the slope unit susceptibility classifications changed in both the information value and ANN models,while only 12% of units changed in the FPR model. The Area Under the Curve(AUC)values for both the ANN and FPR models increased to varying extents. Compared with the comparative models,the FPR model exhibited superior performance(AUC=0.8859,w=1),and the method is characterized by strict mathematical derivation at each step,high interpretability,and reduced reliance on sample data. This provides an effective method to enhance the spatial accuracy of landslide predictions.