Water quality assessment is one of the important means of environmental management. Using water quality evaluation, the water quality and its difference in different areas can be clearly known and the changing trends of water quality in different periods would be also got. Due to natural fuzziness in the water quality and the difference of influence on water quality of each evaluation index, the water quality evaluation may be regarded as a fuzzy problem. A fuzzy comprehensive evaluation model based on entropy weight method (FCE-EW) was built to assess the actual state of underground water quality in this paper. In the FCE-EW, an influence weight of each water quality indictor can be calculated through mining information of the observed underground water samples with the entropy weight method, and the ambiguity of water quality can considered in the fuzzy comprehensive evaluation. The calculation results of case study showed, compared to that of RAGABP and PPC, the evaluation results of underground water by FCE-EW were more objective. So the FCE-EW can be used as a good method to assess water quality in practice.