Abstract: In order to optimize the transfer process and formulation of coating, and improve the comprehensive performance of the transferred coating, an optimized scheme was carried out based an a verified BP neural network model.The evaluation index was the minimum weighted sum of the coating hardness, wear extent and shear strength calculated by genetic algorithm(GA), and the iterative optimization was made on the formulation, vacuum fusion temperature and heat preservation time of the coating. When the weight of the hardness, the wear extent and the shear strength were-0.3, 0.4 and -0.3, the maximum iterative time was 100, and the population was 60, the crossover and mutation probabilities would be 0.40 and 0.01, respectively. The optimization scheme includes keeping WC content at 26.8%，keeping the vacuum fusion temperature at 1071℃ and keeping heat preservation time in 59.7min.After the optimization,the hardness of the composite coating is 63.6HRC, and the wear extent is 101mg only,and the shear strength is 157.2MPa.It shows that GA is scientific and effective in optimizing coating formulation and transfer process, and the comprehensive performance of the composite coating can be improved prominently.

Key words: coating transfer process, composite coating, genetic algorithm, multiobjective, BP neural network, flame spray welding, vacuum fusion