Abstract: Eastern Guangdong Province is characterized by complex geomorphology and abundant precipitation, where heavy rainfall frequently trigger landslide and other geological hazards, posing severe threats to local communities and infrastructure. This study establishes a landslide susceptibility evaluation system for Xiuduan Town, Lianping County, using slope unit segmentation and the information value model. ROC curve validation demonstrates high prediction accuracy with an AUC value of 0.829. Historical rainfall spatial patterns are reconstructed using IMERG precipitation data from the Global Precipitation Measurement mission. Based on the regional geological vulnerability and susceptibility assessment results, scenario-based hazard evaluation is conducted under three rainfall conditions: 24-hour cumulative rainfall (100 mm, 250 mm) and 72-hour maximum cumulative rainfall, calculated using an empirical attenuation formula (first 24 h × 1 + second 24 h × 0.81 + third 24 h × 0.64). Hazard zoning under these rainfall scenarios integrates regional geological vulnerability and rainfall-triggering mechanisms. Comparative analysis shows that the evaluation results correspond well with field investigations, indicating reliable accuracy. The proposed methodology advances quantitative risk assessment for rainfall-induced landslides in complex terrain, offering technical references for early warning systems and disaster preparedness in similar geological settings. Rainfall of varying intensities provides critical guidance for the precision prevention and control of geological disasters and risk avoidance.