In an open and dynamic IoT (the Internet of Things) environment, a common context information model is essential for active things to share common knowledge, reason their situations, and support adaptive interoperability with each other. There have been many studies on the IoT context information models based on semantic technology, but most of them have assumed a static situation based on a service-oriented information model suitable for specific applications of the IoT. In the case of applying their models to an open and dynamic IoT environment, two issues have been observed: Most of the models ignore (a) the mashup of the open-world semantics of context information generated by multiple context sources and (b) the reconciliation of the semantic relationships between multiple context entities under dynamic situation changes. Therefore, in this paper, we propose a context information model that is flexible enough to express complex and diverse semantic relationships between context information generated from a variety of context information sources in the IoT. The main background of this proposal is to propose an adaptive context model that can effectively mash up various context classes that use ontology in open and dynamic IoT environments. In this paper, we also show the effectiveness of the proposed model through an adequate verification model and a practical example.