The thickness variation of drawing part is usually very complicated, which causes the accurate calculation of the surface area of drawing part to be very difficult. Blank optimization of sheet metal forming is often considered as a multi-objective problem. A blank optimization strategy of sheet metal forming process was suggested based on Response Surface Methodology (RSM). Latin Hypercube Sampling (LHS) method was introduced to design the rational experimental samples; the objective function was defined based on crack factor and wrinkle factor; the accurate response surface for sheet metal forming problem was built by Least Square Method; Genetic Algorithm (GA) was adopted in optimization and Pareto solution was selected. The strategy was applied in blank optimization of an auto-part, this method was proved suitable for blank optimization of sheet metal forming.