The detection of insurance fraud poses a significant challenge for the insurance industry. To address this issue, this study leverages the power of XGBoostXGBoost, a scalable machine learning algorithmMachine learning algorithms, to predict fraudulent insurance claims. Utilizing a dataset containing policy details, customer information, and accident-related data, XGBoostXGBoost was employed to develop a robust fraud detection model. Through exploratory data analysis, feature engineeringFeature engineering, and model training, XGBoostXGBoost demonstrated its effectiveness in identifying fraudulent claims while minimizing false positives. Evaluation metricsEvaluation metrics, including precision, recall, F1-score, and AUC, highlighted the model’s performance in fraud detection. Additionally, class imbalance handling and the impact of specific variables on fraud tendencies were explored. This research offers insurance companies a valuable tool for enhancing fraud detection, leading to cost savings and improved security in the industry.