In this thesis, a role selector for the robot soccer is designed and developed based on immune system. In the multi-agent system, it is an important problem to determine an appropriate behavior for the cooperation and coordination among agents. There are many techniques in choosing an appropriate behavior of the agent. As the environment is dynamic, the immune system is suitable than other system. The ROBOSOT is one of the robot soccer category which is a good testbed of the multi-agent system. In the ROBOSOT, a robot has a vision system which finds the position of the ball and the other robots and a communication ability for sharing the information among the agents. To verify effectivness of the proposed method, it is applied to the robots for the ROBOSOT.