Digital twin technology establishes the future development vision for Industry 4.0, and is also an important exploration direction for the Model-Driven Engineering (MDE) paradigm. Because it builds a more flexible and communicative production system through models that spans life cycle, hierarchy and architecture. The standard proposed under the concept of Industry 4.0, the Asset Administration Shell (AAS), provides a syntactic interoperability interface for all assets involved in smart factories. However, there is still a need to fill the gap regarding semantic interoperability, in order to allow efficient interactions between Industry 4.0 components. Ontologies are a good candidate because they provide formal semantics expressed using a knowledge representation language, and in addition, there are many associated mature tools for reasoning and inference. Therefore, we propose a modeling approach that provides semantic interoperability for AAS-based digital twins using ontologies.