As a human-computer interface associated with optimal design of artificial products, it is essential for designers to model adequately their value system having inherently a multi-attributed nature. It aims eventually at developing a use-adaptive intelligent system in manufacturing. With this point of view, in this paper, we have developed a novel design method for multi-objective optimization (MOP) in terms of soft computing. The method has very flexible modeling ability of value function that is likely to fluctuate depending on the decision environment. Actually we have given methods to revise certain inconsistencies in the subjective judgments at the stage of value function modeling. Additionally its formulation allows us to apply any conventional single-objective optimization methods readily for obtaining the preferentially optimal solution. Furthermore, we can employ it regardless of the nature of the model under consideration, i.e., physical model or metamodel. After outlining the solution procedure, effectiveness of the proposed method is examined through an illustrative case study.