The emergence of Industry 4.0 has heralded notable progress in manufacturing processes , utilizing sophisticated sensing and data analytics technologies to maximize efficiency. A vital component within this model is predictive maintenance, which is instrumental in ensuring the dependability and readiness of production systems. Nonetheless, the heterogeneous characteristics of industrial data present obstacles in realizing effective maintenance decision-making and achieving interoperability among diverse manufacturing systems. This paper addresses these obstacles by introducing a hybrid approach that harnesses the power of ontologies, machine learning techniques , and data mining to identify and predict potential anomalies in manufacturing processes. Our work concentrates on designing an intelligent system with standardized knowledge representation and predictive capacities. By bridging the semantic divide and enhancing interoperability, ontolo-gies enable the amalgamation of various manufacturing systems, thereby optimizing maintenance decision-making in real-time. As demonstrated in the experimental results, this approach not only ensures system reliability but also fosters a seamless, integrated, and efficient production landscape.