Auto panel stamping is a complicated plastic deformation process with geometry nonlinear, material nonlinearity and numerous process parameters. The stamping process of a typical auto panel wheel wrap was studied by artificial intelligent optimization and physical experiment. The prediction model of object function was established using artificial neural network. In object function, blank-holder force, drawbead height and fillet radius were selected as the optimized variables and prevention of rupture was considered as the optimization objective. Process parameters optimization was performed with genetic algorithm. The optimized process parameters were used to guide die design and testing, and the result of wheel wrap stamping showed that the forming quality was obviously improved. So the process optimization based on artificial neural network and genetic algorithm is feasible and efficient for auto panel stamping.