This paper presents a collective intelligence approach to collaborative knowledge creation which enables multiple users to collaboratively create knowledge in a systematic and dynamic process. It encourages members to collaboratively express ideas or solutions regarding the complex or challenging issues and submit arguments to support or oppose other members?? ideas. In order to semantically capture and describe the community deliberation and maintain it as a community knowledge-base, a collective argumentation ontology, namely ColaOnto, is proposed by combining and extending IBIS model and Toulmin??s scheme. Moreover, four measures: a degree of argument, degree of individual preference, degree of position, and degree of expertise, are calculated based on the deliberation for evaluating the acceptance level of each position. A position accepted by most members will be selectively integrated as collective knowledge, which is collaboratively developed until the final knowledge is obtained.