TsunamisTsunamis are among the most destructive natural disastersNatural disaster. The accurate prediction of these events is a critical requirement for developing early warning systemsEarly warning systems and effective disaster managementDisaster management. This paper discusses the development of hybrid machine learning modelsHybrid machine learning models for improving the prediction of tsunami events. Machine learning models such as (1) a soft voting ensembleEnsemble learning of random forest and XGBoost, (2) a stacking ensemble of SVM, random forest, and XGBoost, and (3) a hybrid model, called NBTree, which combines naïve Bayes and a decision tree, are used. These models predict the possibility of a tsunamiTsunami through analyzing data related to earthquakes such as magnitude, depth, latitude, and longitude. A comparison of metrics (accuracy, precision, recall, F1-score, sensitivity, and specificity) addresses the most common trade-offs between false positives and false negatives. The findings indicate that hybrid approaches strengthen the prediction accuracy of models and enhance the reliability of accurate forecasting of disaster events. This study demonstrates the potential for integrating advanced machine learningMachine learning techniques into disaster preparednessDisaster preparedness strategies for developing effective AI-driven solutions in managing natural disasters.