Abstract: In order to realize target clustering quickly, accurately and effectively, a target clustering method based on improved label propagation algorithm is proposed. Based on OODA cycle theory, the importance of target nodes is designed to determine the updating order of target nodes and improve the stability of clustering results. According to the similarity of spatial distance between nodes, similarity of network structure and importance of nodes, the label influence value is formed, so as to update the target node label and improve the accuracy of clustering results. An example of air defense system network is given to verify the effectiveness of the proposed method.