Based on genetic optimization algorithm,the diffusion bonding process of T2 copper was studied. Used orthogonal test method and combined with back propagation（BP） neural network and multi-objective genetic algorithm,the temperature,pressure and holding time of the diffusion bonding are taken as input variables and the bonding rate and deformation of the sample were taken as output variables after diffusion bonding.The process parameters of diffusion bonding were optimized,and the corresponding diffusion bonding verification test was carried out.The results show that the appropriate diffusion bonding process parameters of T2 copper are:the temperature is 780 ℃,the pressure is 7.5 MPa and the holding time is 120 min. Under this condition,the bonding rate can reach 95.26% and the deformation is 0.166 mm.The above-mentioned process parameters are used to manufacture the parts of the microchannel heat exchanger,and the deformation in the thickness direction is 0.162 mm.After ultrasonic C-scanning,the bonding condition is good, and the sealing and design requirements are satisfied after the pressure-proof and leak-proof detection.