Abstract: Considering the characteristics of anti-terrorism task and the platform-based and data-based development trend of command information systems, it is urgent for the command system to have the initiative and accurate command information recommendation ability. Collaborative filtering algorithm is a representative personalized recommendation technology with a good application prospect. However, when collaborative filtering relationship is established between commanders and command elements, collaborative filtering algorithm faces the problems of data sparse and cold start, which result in ineciency and inaccuracy. This paper proposes a multicollaborative filtering algorithm for anti-terrorism tasks. The algorithm first performs project-based collaborative filtering by combat type on the command elements, and then combines the cohesion subset analysis into user-based collaborative filtering to mine the similarity between the commanders and command elements under specific combat types, thus to achieve precise recommendations. The experimental results show that the proposed algorithm can be applied to command information systems of anti-terrorism tasks effectively and improves the recommendation efficiency and accuracy.