This research paper explores the use of the AdaBoost algorithm for predicting residential habitation status in the aftermath of the Mount Merapi eruption. Using a dataset from the Rehabilitation and Reconstruction Task Force, with 2516 instances and 11 attributes, the AdaBoost model was trained and evaluated. The model demonstrated a robust performance with an accuracy of 92%, though it struggled with correctly identifying all 'No Habited' instances. These findings underscore the potential of machine learning algorithms in disaster management, particularly in optimizing resource allocation and expediting recovery times. Future research should aim to improve the model's performance on imbalanced datasets and explore the incorporation of temporal dimensions for more dynamic and accurate predictions.