Predictive maintenance plays a crucial role in improving the reliability and efficiency of industrial systems byenablingearly fault detection and reducing unplanned downtime. This study presents a machine learning-based predictive maintenanceframework using Decision Tree and Optimised Random Forest algorithms, evaluated on the Commercial ModularAero-Propulsion System Simulation (C-MAPSS) dataset developed by NASA. The Decision Tree model is implementedas abaselineclassifier, achieving an accuracy of 86.19%. To enhance predictive performance, an optimised RandomForest model is developedusing ensemble learning techniques, resulting in improved accuracy of 88.54%. The models are evaluated usingaconfusionmatrix, receiver operating characteristic (ROC) curve, and k-fold cross-validation. The results demonstrate that theRandomForest model provides better classification accuracy, improved generalisation, and reduced misclassification comparedtotheDecision Tree model. The study highlights the effectiveness of ensemble learning techniques for predictive maintenanceincomplex engineering systems. Keywords: Predictive Maintenance, Machine Learning, Decision Tree, Random Forest, C-MAPSS, Classification, EnsembleLearning.