Every product goes through many stages of the life cycle from manufacturing through usage to utilization. Some stages can change its properties, which in turn changes its description. To describe the product and its life cycle many ontologies have been created with varying levels of detail. Ontologies usage at different product lifecycle management (PLM) stages provides a better match to these stages since they provide properties of the product important only for this stage. During the transition between stages, the data should be integrated across all systems in manufacturing domain to provide semantic interoperability. Therefore, an issue arises of matching descriptions presented with ontologies of lifecycle stages. This is especially critical if ontologies for different stages are created by various specialists (for example, designer, technology engineer, retailer, maintainer, etc.). The paper proposes the method of matching ontologies for the formation of a common PLM ontology based on the automatic matching of ontologies referred to PLM stages. It allows to overcome heterogeneity and ensure interoperability in the process of tracking the product through the PLM stages. The matching process is based on the identification of common concepts by which ontologies will be combined into one. To identify common concepts, the ontology matching method is used, based on a combination of a context-based matching with neural network to find similarities of concepts (name, characteristics names and their string values) and study the ontology structure to identify common design patterns.