Services generated by various Internet of Things (IoT) devices are in heterogeneous contexts, hindering users from efficient consumption of those services and hence development of IoT. This paper addresses the problems in user-device interaction. A user-device interoperability framework, Tabdoc approach, is hereby established to implement semantic interoperability between IoT devices and users. In this approach, a divide-and-conquer strategy is used; a semantic document representation method for message modeling is devised to guarantee consistent message understanding among different contexts, and a semantic extraction algorithm and a semantic interpretation algorithm are proposed to implement automatic cross-context message processing. The experimental results show promising performance improvements over state-of-the-art methods.