In order to realize fuzzy comprehensive evaluation of the information warfare of air defense and clear off the redundant data, in indexes membership, which are not useful for goal classification, and based on data mining of entropy, mining knowledge information about object classification hidden in every index, the relation between object classification and index membership is affirmed, the redundant data in index membership for object classification are eliminated by defining distinguishable weight, and valid values are extracted for computing object membership. A case of the information warfare of air defense is adopted in this paper. Empirical results show that: In order to clear the interference of redundant data, the membership conversion model can be used as a valid method with multiple indexes attributes. By using the method based on data mining of entropy clear the interference of redundant data in fuzzy evaluation is cleared off and many multiple indexes decision-making problems are solved.