<p>Gas turbine engines experience degradation in rotating components due to fouling and erosion, impacting blade surfaces and decreasing performance by 70–85%. Timely prediction of such wear is critical to maintain engine health, ensuring reliable and efficient operation. These degradations can reduce thrust, high exhaust gas temperature, and increase fuel consumption. Diagnosing the degradation using the data-driven approach aided by the supervised learning algorithm can provide vital results to recover the gas turbine engine. This application is useful for detecting faults in turbomachinery through the engine’s operational data to avoid further damage to the engine’s defective component(s). This study applies a data-driven approach to predict degradation on a simulated model of an open-access dataset of a high bypass ratio unmixed turbofan engine from the NASA database. Three supervised learning algorithms (decision tree, random forest, and support vector machine) are trained, tested, and validated to classify between healthy and degraded engines with degraded component(s) and type of degradation(s). The final model was trained on 336,000, validated, and tested on 112,000 sample data points each; however, the comparison of model accuracy was drawn at three different sizes of training, validation, and testing datasets to study the effect of the dataset on the accuracy of the classification algorithm. The model once trained can be used to predict the degradation of engine components for the new input data, the trained classifier can be utilized in engine test facilities to predict the engine degradation at runtime during engine operation to salvage the engine. The study is concluded by comparing these algorithms based on accuracy and computational cost for training the classifier. The support vector machine with one versus one classification approach has provided the most accurate results for degradation classification with an accuracy of more than 90% among all the tested datasets.</p>