<p>Urban flooding presents serious threats to contemporary cities, driven by the joint effects of climate change and accelerated urbanization. This study presents a model for assessing urban flood susceptibility using Bayesian Network, designed to quantify uncertainties and capture interdependencies among influencing factors, with Beijing as a case study. By integrating climatic, topographical and hydrological, and socio-economic data, the topological structure of flood influencing factors is constructed using the Peter-Clark algorithm combined with domain knowledge, followed by parameter learning through maximum likelihood estimation. A flood susceptibility map is produced, identifying high-risk areas chiefly located in central urban zones with flat terrain, extensive impervious surface coverage, and dense population. The model’s validation, including ROC curve analysis and accuracy metrics, demonstrates a high degree of reliability. The results reveal three key flood-inducing factors chains spanning three dimensions. Sensitivity analysis further emphasizes the dominant roles of rainfall, land use, and socio-economic factors in driving flood susceptibility, offering insights into flood propagation pathways. These findings provide valuable guidance for targeted flood mitigation while informing strategies for sustainable urban planning and resilient infrastructure development.</p>