Hot forming-quenching integrating process of aluminum alloy sheet has been developed to improve formability and overcome thermal distortion by integrating hot stamping and quenching in one operation, which is significantly beneficial for the application of aluminum alloy components in vehicle lightweight industry. In this novel technology, processing parameters are interrelated and interacted. To obtain appropriate forming parameters, thermo-mechanical finite element (FE) model using a unified viscoplastic damage model is set up to predict the formability of a complex-designed 6xxx aluminum alloy B-pillar. A Back Propagation (BP) neural-network combined with multi-objective genetic algorithm (GA) method is adopted to optimize the key process variables including blank temperature, stamping speed and die clearance during hot stamping process. After optimization, the thinning and thickening are reduced to 13.0% and 10.0% by comparing with the initial 56.5% and 14.2%, respectively. In addition, a successful hot stamping B-pillar with satisfactory mechanical performance and excellent forming accuracy is achieved experimentally by using the optimized parameters. It indicates that the finite element model can simulate the hot stamping process accurately and that the optimization method utilized in this paper is feasible and effective.