Properly evaluating the beach wetland ecological services value will contribute to the awareness of beach wetland protection for people, and provide basis for the planning of beach wetland reclamation ecological compensation criterion. The scientific nature of traditional classification calculation method of ecological services value is disputing, and the common CVM data handling techniques have some limits on the recognition of investigation results analysis. This research establishes a fuzzy evaluation model for beach wetland ecological services value based on CVM method. The prices "cheap", "moderate" and "expensive" are three fuzzy subsets for specific domains of discourse. This paper highlights the comparison of different membership grades for different prices on different fuzzy subsets, establishes a traditional fuzzy statistical model, weighted fuzzy statistical model and three-phase model according to non-uniform discrete membership grades, weighted discrete membership grades and successive membership grades based on the interval data of CVM, and evaluates the annual ecological services value per unit area of Hangzhou Bay National Wetland Park. The results showed that:the unit prices obtained through 3 methods are 10.28 Yuan/m², 10.38 Yuan/m² and 9.76 Yuan/m², which have high uniform degree and similar with domestic beach wetland values. The interval of moderate price for wetland ecological services is[5.46,15.33] which is obtained when dividing the model through three phases and the subordinate function values of each "moderate" price in the moderate price interval are higher than the subordinate function values of "cheap" and "expensive". Innovations of this study are as follows. First, the innovation value is the rationality of fuzzy mathematical model analyzed through the conceptual "fuzziness" of "ecological value", and find the representative price from the maximal onset principle of membership grade, which is different with the data central tendency principle of traditional CVM data processing; second, the "rugby" type weighted statistical model established can overcome the irrationality of the uniform weighting to interval data of the traditional fuzzy statistical model, and the results of smoothness and degree of fitting on the membership function are better than the traditional model; third, three-phase model fits the membership functions for three subsets which could compare the membership grades difference of different fuzzy subsets with same prices. The established model is reference significance for the evaluation of ecological resources value based on CVM.