In the near future, homes are envisioned to be equipped with numerous intelligent communicating devices. Such smart home needs to exhibit highly adaptive behavior to meet the inhabitants changing personal requirements and operational context of environment. To achieve this, smart home application should focus on the inhabitant?s goal or task in diverse situation, but not the various complex devices and services. This paper proposes a context-dependent task approach to meet the challenge. The most important component is task model which provides an adequate high-level description of user-oriented tasks and their related contexts, and in such model multiple entities can easily exchange, share and reuse their knowledge. An OWL-based ontology to hierarchically model context-dependent task is presented, which facilitates sharing and reusing of smart space knowledge and logic inferences. The conversion of OWL task ontology specifications to the First-Order Logic (FOL) representations is described. Finally, the performance of FOL rule based deducing in terms of task number, context size and time is evaluated.