A TOPSIS missile target threat assessment algorithm based on gradient fuzzy is proposed for the problem of missile target threat assessment. By using the TOPSIS（technique for order preference by similarity to ideal solution） method to calculate weight values and combining gradient fuzzy calculation with the fuzzy matrix， the comprehensive weight values of target threat assessment indicators can be obtained， which can solve the problem of weight deviation caused by large differences in the magnitude of data indicators for various factors， such as large or small deviation. The evaluation results of typical cases indicate that the proposed improved algorithm overcomes the problem of weight difference and effectively enhances the credibility of missile target threat assessment.