During the manufacturing product life cycle, an increasing volume of data is generated and stored in distributed resources. These data are heterogeneous, explicitly and implicitly linked and they could be structured and unstructured. The rapid, exhaustive and relevant acquisition of information from this data is a major manufacturing industry issue. The key challenges, in this context, are to transform heterogeneous data into a common searchable data model, to allow semantic search, to detect implicit links between data and to rank results by relevance. To address this issue, the authors propose a query system based on a graph database. This graph is defined based on all the transformed manufacturing data. Besides, the graph is enriched by explicitly and implicitly data links. Finally, the enriched graph is queried thanks to an extended queries system defined by a knowledge graph. The authors depict a proof of concept to validate the proposal. After a partial implementation of this proof of concept, the authors obtain an acceptable result and a needed effort to improve the system response time. Finally, the authors open the topic on the subjects of right management, user profile/customization and data update.