A new TOPSIS model which based on the gray correlation coefficient matrix was presented in order to overcome the deficiency in more indicator of decision-making process. The method took the gray correlation coefficient matrix between primitive data sample and ideal scheme as new decision-making matrix,and then TOPSIS was used to evaluate and arrange all schemes. This method overcame the traditional TOPSIS shortcoming of excavating the inherent law of the data difficultly. Meanwhile, AHP and entropy methods were used to decide the weight of every decision-making index. This overcame the subjectivity or objectivity of traditional entropy decision methods. The new evaluation model was used in water quality in this paper. The result of application adapt to the evaluation of water quality.