Community detection is a hot topic in social network analysis, and many existing studies use graph theory analysis methods to detect communities. This paper focuses on detecting absolute fair maximal balanced cliques in signed attributed social networks, which can satisfy ensuring the fairness of complex networks and break the bottleneck of the "information cocoon ".