Summary: Reinforcement learning control is one of the important approaches in intelligent control. The immune system can recognize the invaded antigen and eliminate it rapidly, and is more stable when the antigen invades again. This is the primary-secondary response mechanism of the immune system, and can be regarded as a process of reinforcement learning. Based on this mechanism, a novel reinforcement learning intelligent controller (RLIC) is presented in this paper. The RLIC has the abilities of learning, memory, and evolution. The control system has no control antibodies (CABs) when the control error appears for the first time. The RLIC can learn and produce the CABs automatically ducing the period of eliminating the control error, and the CABs store the final changing control output after the error is eliminated. When the control error appears again, the RLIC can eliminate it rapidly and stably by using the stored CABs and combining the conventional control algorithm. After the control error is eliminated, a new CAB is produced and stored. Repeating the above process, the RLIC’s learning ability and system response become stronger and stronger. Consequently, the control performance of the RLIC can be improved. In order to examine the effectiveness of the RLIC, the simulation experiments are carried out by choosing a second-order plant with a time delay and a third-order plant. Simulation results demonstrate that system response ability and stability of the RLIC are better than those of the conventional PID controller.