This contribution introduces a framework for the fault detection and healing of chemical processes over wireless sensor networks. The approach considers the development of a hybrid system which consists of a fault detection method based on machine learning, a wireless communication model and an ontology-based multi-agent system with a cooperative control for the process monitoring.