Abstract: Elevator group control system is one typical hybrid dynamic system.It includes continuous states，discrete states and complex human logic strategy.In this paper，cellular automata theory is applied in elevator group control system modeling and one Hybrid elevator group control system model is built up.The proposed model can describe the continuous states and the discrete states accurately and can combine elevator movement module with management module effectively.This hybrid elevator group model is simple and can be applied in many different fields.This paper also proposes a novel elevator group control dispatching policy with artificial immune algorithm.Artificial immune dispatching policy with high variation cloning can avoid going to the system local optimum and can improve optimal searching speed.Simulation results demonstrate the feasibility and superiority of the proposed hybrid elevator system model and artificial immune dispatching policy.