Abstract: A kind of fast and elitist multi-objective genetic algorithm (nondominated sorting genetic algorithm -II) was presented to solve high dimension and multi-modal optimal problems. T His fuzzy information could be converted into a mathematically well-structured problem based on fuzzy optimal theory. And the improved crossover operator of NSGA-II was applied to obtain the optimal solution. According to the test results on a typical test function and an application on the structural fuzzy multi-objective optimization of three-bar truss, more reasonable distributed solutions could be obtained and the diversity of the solutions could be maintained. It provides beneficial references for engineering application of fuzzy multi-objective structure optimization.