Abstract: For the multi-attribute decision-making problem with preference information on alternative，in which the attribute weights are known incompletely，this paper proposes a new decision method.After solving the optimal problem of minimizing the maximum deviation between the subjective and objective preference，the vector of attribute weights are reasonably determined.The method can overcome the drawback that there is no guarantee that the deviations between the subjective and objective preference for each alternative are all minimized in the existing methods.Finally，the applied example of product development proves that the approach is feasible and practicable.

Key words: multi-attribute decision-making, incomplete information, preference, min-max

摘要： 针对部分属性权重信息且对方案有偏好的多属性决策问题，提出了一种新的决策方法。该方法通过求解最小化最大的主、客观偏差的数学优化模型，较合理地确定属性的权重向量。克服已有方法中不能保证每一方案的主、客观偏差都最小的问题。产品开发应用实例表明方法的有效性和实用性。

关键词: 多属性决策, 不完全信息, 偏好, 最小最大