Context-aware systems in a smart space environment must be aware of the surrounding contexts and adapt to changing contexts in highly dynamic environments. Data managements of contextual information are different from traditional approaches because of the contextual information are dynamic, transient, and fallible. Consequently, capabilities to detect context inconsistency and maintain consistent contextual information are two key issues for managing contexts. We propose an ontology-based model for representing, deducing, and managing consistent contextual information. In addition, we use ontology reasoning to detect and resolve context inconsistency problems, which will be described in a Smart Alarm Clock scenario.