Operational knowledge is an important factor to improve the autonomy of unmanned systems. How to store and represent operational knowledge semantically and normatively, so that unmanned systems can autonomously accomplish task planning according to the high-level order of humans, carry out re-planning, and make decisions has been a new topic in the autonomous unmanned system field. This paper reviews several typical robot ontologies, as well as modeling and representation methods of uncertain information, proposes the concept of Ontology-based Task Planning(OTP) for the unmanned system, and introduces the principles of ontology construction. Finally, the application of ontology in unmanned combat systems are discussed.