As computing technology is becoming more tightly coupled into dynamic and mobile world of the Internet of Things (IoT), security mechanism is more stringent, flexible and less intrusive. Scalability issue in IoT makes Identity Management (IdM) of ubiquitous objects more challenging, and there is a need of context-aware access control solution for IdM. Confronting uncertainty of different types of objects in IoT is not easy. This paper presents the logical framework for object classification in context aware IoT, as richer contextual information creates an impact on the access control. This paper proposes decision theory based object classification to provide contextual information and context management. Simulation results show that the proposed object classification is useful to improve network lifetime.