Abstract: [Objective] This paper aims to test the theoretical validity of CVM about evaluation of recreation value in forest scenic areas.[Method] The three regression models were established to study theoretical validity of CVM based on improvements in the effectiveness of experiment, which taking National Forest Park of Fuzhou as a case study.[Result] 1) The annual recreation mean value per person is 37.930 yuan and the total recreational value is 1.8 billion yuan in the National Forest Park of Fuzhou, which indicates that it is lower than large scenic spots in the recreational value, because of the structure of the scenic forest recreation value is relatively single. 2) The theoretical validity of CVM in the evaluation of forest scenic area is good. The result is consistent with the economic theory and it can give a reasonable explanation in real life. Whether eliminating sample selection bias of willingness to pay or for the full sample, the amounts of willing to pay are subjected to the constraints of the income level. Amounts of willingness of payment are easier constraints than the willingness to pay by the feature of tourists and evaluation object due to flattery deviation. Urban residents have better quality recreational alternatives for the higher life style, which results in a lower willingness to pay or amounts of payment. Marginal willingness and amounts of pay are the most strong among occupations which states clearly the most true willingness to pay probably about the persons in public sector. The marginal willingness and amounts of payment are negative compared with others; farther from the visitors from the forest park for retirees due to restrictions by the income. Tourists' willingness and amounts of pay decreases as the distance increasing from National Forest Park of Fuzhou.[Conclusion] The CVM can be introduced into the evaluation of recreation value about forest scenic spot based on the improvement of effectiveness in terms of theoretical validity, but the other criteria of validity for CVM should also be fully considered.

Key words: CVM, forest scenic area, recreational value, theoretical validity, Logit regression model, Tobit regression model