Predictive Maintenance (PdM) in the manufacturing sector is crucial for monitoring structure failures and scheduling the Maintenance of machines. This research aims to use machine learning (ML) methods to build and validate precise models for predicting Maintenance and avoiding failures in industrial machinery. It also focuses on evaluating the efficacy of different ML models in predictive maintenance within the manufacturing sector. The study seeks to identify which models can most effectively predict equipment failures, with a detailed comparison of the performance of ANN, LR, and DT models based on various statistical metrics. Their performance was assessed utilizing F1-score, accuracy, recall, and the area under the curve. The results recommend that the Decision Tree (DT) model offers the best predictive performance with an F1-score of 0.982, an accuracy of 0.978, a recall of 0.969, and an area under the curve of 0.980. The most significant influencing features for predicting failure mechanisms were the mean of rotation, pressure, and error 4 count through twenty-four hours. The study suggests that ML can improve maintenance estimates, enhance equipment reliability, and realize the best realizable value in the manufacturing sector.