 This study evaluates landslide susceptibility using Bayesian Optimized Convolutional Neural Network, BOCNN, Bayesian Optimized Random Forests, BORF, and Particle Swarm Optimization Support Vector Machines, PSOSBM, models in a 10 km area surrounding the 7 isobar of the JMA earthquake. The BOCNN model outperforms the other two models, with an accuracy of 0.9535, AUC value of 0.9529, and FR value of 14.9 in the high susceptibility area. The results suggest that the BOCNN model is a significant contribution to disaster prevention and mitigation measures of local governments. This article was authored by Yunlong Deng, Shaoqing Zwa, Yang Fali, and others.