Windstorms are becoming increasingly devastating and frequent climate-related disasters, affecting both lives and economic assets. With advancements in Artificial Intelligence (AI), disaster studies have made significant progress in predicting and modeling risks, particularly in data-scarce regions. This study employs a multi-method approach, combining geospatial data, Machine Learning (ML) models, and Explainable AI (XAI) algorithms to predict windstorm risk in urban Bauchi. The analysis was conducted using 4651 locations, evenly distributed between areas with and without a history of windstorm destruction. Three ML models—Random Forest (RF), XGBoost, and Support Vector Machine (SVM)—were applied to predict risk based on nine hazard and vulnerability conditioning factors. The predictive capacity of the models was evaluated through various metrics and validated using ROC-AUC scores. While all models demonstrated strong windstorm prediction capabilities (ROC-AUC > 0.92), XGBoost and RF showed superior accuracy, with ROC-AUC scores of 0.98 and 0.97, respectively. Both models effectively identified the worst-hit areas from recent windstorm events as high-risk zones. Additionally, XAI was used to examine the impact of the nine conditioning factors on windstorm risk. Both RF and XGBoost revealed consistent trends, identifying housing density, NDVI, and population density as the most influential factors. Densely populated areas were found to be more vulnerable to windstorm hazards, likely due to structural vulnerabilities and increased exposure. These findings highlight the strong connection between urban densification and windstorm risk and exposure, underscoring the need for achieving urban resilience in Bauchi through integrated spatial planning initiatives.