Abstract: Took a complex thick-wall part as the research subject, established the finite model. Analyzed it’s maximum warping deformation under the initial process by CAE. Build the plastic part warpage quantity optimization model, based on the BP neural network combined with genetic algorithm to optimize the processing parameter of the model, optimal process parameters were showed the maximum warping deformation was 0.2313mm, compared to the initial maximum warping deformation which obtained in the initial process by CAE was reduced by 21.53%, improved the quality of the plastic parts and satisfied the requirements of the assembly parts. Research results further confirmed the validity and reliability of the proposed BP-GA based optimization method for plastic injection forming.