This paper examines optimization techniques for parametric design based on statistical, mathematical, and heuristic models. A hybrid method is then proposed based on the strengths of several of these approaches. The new method, called multi-target parametric design method, combines Taguchi, methodology, conventional (mathematical) optimization methods, and MANOVA statistical technique. It can be used to optimize composite utility functions, while satisfying the quality requirements. The method involves three main steps. First by applying Taguchi's method, the optimal quality is determined with a certain set of control variables. Second, these variables are transformed into variable constraints through a relaxing process. Finally, the optimization problem is solved for other design targets under these constraints, using conventional mathematical optimization methods or the Taguchi method.