Abstract: Stated preference method is usually used to evaluate the non-market value of environmental goods which includes contingent valuation method (CVM) and choice experiments (CE). In this paper, stated preference method was adopted to evaluate the non-market value of Sanjiang Plain wetland. A willingness to pay (WTP) evaluation model of stated preference method was constructed based on the random utility theory. The average WTP of CVM and CE was obtained, respectively. The average WTP elicited by CE was 379 yuan per year, and the marginal WTPs of different selection properties including water conservation, wetland area, natural landscape and biodiversity were114.00, 72.55, 59.55 and 37.09 yuan per year, respectively. Meanwhile, the average WTP elicited by CVM was 134 yuan per year. The influence of factors on WTP was analyzed and reasons for protest responses were discussed. Results showed that the respondents’ WTP elicited by CE was signi-ficantly higher than that by CVM, and respondents’ socio-economic attitudes such as level of education and personal annual income had a significant positive impact on respondents’ WTP. There were no significant difference in the reasons of protest responses between CVM and CE. Besides, respondents’ multiple attributes and multiple levels analysis could be carried out by CE and the WTP of wetland’s selection attributes could be calculated. Therefore, CE had the better ability of revealing respondents’ preference information than CVM and its assessment results were more close to the actual value.