The ever increasing demands for improved performance in engineering design from an overall perspective requires a sophisticated automated design generation strategy that can effectively and efficiently incorporate concepts of uncertainty, quality and robustness into design. Engineering design optimization approaches that typically require modeling the problem in a single objective form become inadequate to address these multiple set of requirements. Instead, this research presents an alternative robust multiple-criteria based design approach. In this approach, decision models are formulated using utility functions that quantitatively represent designers’ subjective preferences and optimal solutions are then determined in a statistical exploration based iterative design of experiment set-up. The resulting Trade-off Based Robust Engineering Design (TRED) method is illustrated with the aid of an example design problem. The results indicate that the TRED approach can be expected to address some of the important issues associated with the engineering design of problems including tolerances, quality control and uncertainties.