The digital twin workshop is a new workshop operation paradigm that enables precise decision-making by fusing virtual and physical space. As a complex manufacturing system, the digital twin model of the workshop must integrate models from different domains in order to provide personalized services. The interoperability of multi-domain models and the dynamic update of parameters become obstacles. In this paper, a knowledge graph (KG)-based multi-domain model integration method for digital twin workshops is proposed. The multi-domain model integration architecture based on KG is consisted of model element, model ontology, model data, semantic integration and network connection. Then, the KG of multi-domain model for design, manufacturing and simulation is constructed through ontology modeling and knowledge extraction. On this basis, multi-domain model integration is realized through semantic inference and knowledge query. The model parameters are updated through le exchange during the dynamic simulation. Finally, multiple scenarios in the subassembly workshop for hull construction are used to verify the e cacy of the proposed method. During the assembly and welding of hull parts, the integration of the product model, equipment model, and simulation model is realized, which assists in meeting the service requirements of multiple business scenarios.