This study investigates the use of predictive maintenance and condition-based monitoring (CBM) strategies for maritime drive systems through the application of machine learning techniques. These systems are critical to vessel operations, where performance degradation or unexpected failures can result in costly downtime, increased operational expenses, and safety hazards. Key operational parameters—including torque, rotational speed, temperature, pressure, fuel consumption, and state indicators—are continuously monitored to assess system health and performance. By leveraging both historical and real-time data, machine learning models are trained to detect anomalies, predict failures, and support system optimization. Regression algorithms are evaluated using performance indicators such as Mean Squared Error (MSE) and the coefficient of determination (R²), ensuring accurate and robust predictions. Through the analysis of interdependencies among system variables, the models uncover patterns associated with wear, inefficiency, or impending faults. This predictive framework enables proactive maintenance decisions, minimizes unscheduled repairs, and improves the reliability and operational efficiency of maritime propulsion systems. The outcomes offer practical insights for enhancing performance management and promoting higher safety standards in the maritime industry.

Predictive Maintenance, Condition-Based Monitoring (CBM), Maritime Drive Systems, Machine Learning, Anomaly Detection, Failure Prediction, Propulsion System Monitoring.