Hydraulic systems are employed in many different industrial domains; their function and level have been advanced. Furthermore, system management and maintenance include condition monitoring via Internet of Things (IoT). In order to improve anomalies and flaws in each hydraulic system component (HSC), meaningful features were extracted and selected and then used machine learning (ML) classifier in this research. In predetermined portions, defected data are gathered from IoT sensor data that were grouped in to clusters. By using a cluster, the shape and density properties were extracted. For every HSC, associated features were selected from 2335 extracted features by using the modified bald eagle search (M-BES) method, and these features were then used for model learning. LightGBM was utilized to distinguish between normal and abnormal circumstances and determine the evaluation metrics such as accuracy, positive predictive value (PPV), sensitivity, specificity, and f1-score for HSC. When compared accuracy performance of valve condition, hydraulic accumulator and internal pump leakage, data to support vector classifier (SVC), decision tree (DT), and random forest (RF), the proposed model achieved a 0.964 performance.