Shift schedule determines the gear-shift timing of stepped automatic transmissions that directly affects the power performance, fuel economy, and emission of vehicle. Under the circumstance with ever increasing concerns about the issues of energy saving and environment protection, an optimized shift schedule that could make the powertrain system provide sufficient power with minimum fuel consumption and less emission has been becoming a very critical factor to improve the overall performance of vehicles equipped with stepped automatic transmission. Traditional methods usually only consider a certain unique performance index, such as power or economical performance, in a certain shift schedule, i.e., using one unique performance as optimization objective for one shift schedule. In this study, the performance models of engine including indices of torque, fuel consumption, and emission have been built respectively using artificial neural network through training of the test data of engine, all three types of performance indices including those of power, fuel economy, and emission have been considered together in one uniform evaluation function using a performance weighted method to reflect the preference of driver, and a genetic algorithm (GA) has been applied to optimize the shift schedule that takes the proposed evaluation function as optimization objective, and the throttle opening and velocity as design parameters. Simulation results show that the proposed optimization method for the shift schedule is applicable, and can improve the comprehensive performance of vehicle equipped with stepped automatic transmission effectively.