Abstract: Taking the characteristics of multi-constraint QoS (quality of service) routing in NGI (next generation Internet) into account, a QoS multicast routing algorithm based on MEC(mind evolutionary computation) and game theory is presented. It introduces the membership functions to evaluate the adaptability of candidate edges to users requirements for QoS; makes sure of whether the network provider has been in Nash equilibrium with users in respect to the utility on edges through gaming analysis; constructs a multicast routing tree via MEC so as to enable users' QoS requirements to be satisfied and the utility of both network provider and users to reach or approach Pareto optimum when both are in Nash equilibrium. Simulation results showed that the proposed algorithm is feasible and effective.