Abstract: Considering the decision making problem of emergency supplies dispatching for major natural disasters, we propose a multi-objective optimal emergency dispatching model considering road reliability after disasters. Taking the minimum running time of the longest vehicle, the maximum reliability of the minimum vehicle route, and the minimum unsatisfactory degree of system materials as objective functions, we use the second-generation non-dominated sequencing genetic algorithm (NSGA-Ⅱ) and weighted genetic algorithm (GA) to solve the functions. The results are compared and analyzed. A simulation was conducted, taking Wenchuan Earthquake as an example, and the results show that the optimal solution of NSGA-Ⅱ is better than that of the weighted GA algorithm in multi-objective emergency dispatching problem, and the convergence speed is faster, which verifies the validity of the emergency dispatching model.