As the river basin tends to balance both water regulation and power generation during the actual dispatch process, it’s essential to carry out researches on multi-objective optimal dispatch of cascade reservoirs, which is a complicate multi-objective problem. To solve this problem, an improved multi-objective non-dominated sorting genetic algorithm based on NSGAII is proposed. The proposed algorithm adopts a multi-operator strategy to escape premature convergence and an elite operator strategy to enhance the convergence rate. To improve its performance, case study has been done to Hanjiang river basin with 5 reservoirs and 23 water users. The results show that the improved multi-objective evolutionary algorithm can obtain a reservoir dispatching plan that can more effectively coordinate the contradiction between water supply and power generation, and can maximize the comprehensive benefits of various water conservancy projects in the Hanjiang river basin. Simulation results demonstrate that the proposed algorithm is an effective candidate in multi-objective optimal dispatch of reservoir group.