Abstract: This paper takes into account the properties of web service, the diversity of quality evaluation as well as the description of the vagueness of the language.First, it converts the quantitative and qualitative evaluation indicator of web services into the triangular fuzzy number, and constructs a group decision making function based on the full use of the triangular fuzzy number nature, realizing normalization of indicators.On this basis, we propose a fuzzy-multiple attributive group decision making algorithm based on priority matrix sorting to implement comprehensive evaluation for web services.Finally, this method is proved feasible and effective in the evaluation of the web services by means of instance analysis.