Having cooperation between multiple autonomous devices against one task is difficult due to each device having their own decision management based on self-deterministic protocol. Within the self-deterministic protocol, a formation management task should be considered along another task in order to provide cooperation and consideration between the operating autonomous devices. In this research, a compound learning control system for formation management of multiple control agents is proposed by managing coordination between multiple autonomous agents along with other tasks simultaneously in an operation. A series of simulation based on an autonomous robot was conducted to evaluate the effectiveness of learning through compound knowledge for providing consideration among achieving goals or coordination configuration against partner robot. The proposed system was able to provide consideration in coordination among operating partners in a task of achieving goal.