The paper analyzes the pros and cons of classical gray relational analysis on solving multi-target and multi-attribute problems, and presents an improved gray relational model to assess the battle damage of surface warship formations. First, with an entropy method and analytic hierarchy process, the subjective and objective weights of the index attributes are determined, respectively, and then they are combined with the product method to obtain the combined weight. Second, to make full use of the information carried by the index attributes, a polygon area formed by the adjacent points of the reference sequence curve and alternative sequence curve is used as gray relational coefficient. Third, by combining with the combined weight and the gray relational coefficients, an improved gray relational degree is thus determined. Finally, the numerical results illustrate that the battle damage assessment is generally consistent with the assessment results of experts, and the achieved result will provide important information to make a correct decision for carrying out a next strike.