Abstract: In view of the uncertainty of evaluation indexes of water quality and the fuzziness of water quality standard, a fuzzy connection degree model of water quality evaluation was constructed based on set pair analysis and fuzzy analytical hierarchy process. First of all, the index values of water samples were preliminarily classified by calculating the hierarchical connection degree of each evaluation index value. Then the comprehensive degree of connection between samples and water quality standard was calculated. Finally, water quality grade was judged by confidence criterion. To highlight the contribution of different evaluation indexes, entropy method and super weighting method were introduced. Then the weights were combined based on ideal point method, by which the index weights were more reasonable. This model was applied to the evaluation of the fishery waters of Minjiang River, and the result was compared with those from gray classification method, synthesis index method and single factor evaluation method. The results obtained by the proposed model were closer to the real situation, and hence are reliable.

Key words: set pair analysis, classification of indexes, fuzzy connection degree, combination weight, water quality evaluation