This article presents a flying social network (FSN), which is a kind of flying ad hoc network (FANET) that implies human attributes and relationships. By the accurate analyzed social relationship of unmanned aerial vehicle (UAV) users, the social relationship can act as an essential basis to overcome the information propagation issue in FANET, including long delay and low delivery ratio. Specifically, this article proposes a comprehensive architecture of FSN and describes each of its function modules. Benefitting from social guidance, the devised schemes can improve the quality of service (QoS) of UAV users, enlarge their service coverage area, enhance security, and save power consumption. Then, we introduce a novel social-based routing scheme and design a new caching scheme to reduce the response delay and increase the response probability in FSN. In addition, we highlight open issues and discuss future works of FSN. Finally, a case study is completed by using simulations to demonstrate the effectiveness of the social-based method.