Gujiao mining section in Fenhe River is a main recharge area of groundwater source of Taiyuan, and the pollution in the section is more serious.The tributaries in the section is one of several pollution sources. The paper applied the fuzzy pattern recognition model with the method of correlation coefficient to determine the index weights so as to assess the water quality of the tributaries of Gujiao mining section in Fenhe River comprehensively.The results showed that the level characteristic value of water quality of nine tributaries is between 1 and 2,and the water quality of the section is better.The water quality level in Dachuan, Shizi and Tianchi river is approximate level Ⅰ, and that of other 6 tributaries is level Ⅱ.The evaluation result is reasonable and objective,which can provide academic reference for the pollution control and comprehensive management of water resources in Gujiao section of Fenhe river.