Abstract: Recent studies on the invalidity of contingent valuation method(CVM)have mainly focused on two aspects. First, the issue of scalability of the environmental goods’ non-use value due to the survey respondents’ lack of familiarity with the value of those environmental goods, and secondly the biases and errors arising from the respondents’ willingness to pay(WTP)or willingness to accept(WTA)were more likely to be an expression of preference rather than the real intention to purchase in the real market. CVM valuation on forest environmental service could be improved in two ways. Firstly, to improve survey respondents’ familiarity with forest’s environmental value by focusing on the increase in extrinsic information rather than the increase in intrinsic information. The use of photograph showing the background and the description of environmental goods, conducting face to face interview, together with the use of visual-simulation technique to show differential changes in forest ecological system, will increase survey respondents’ degree of perception on the environmental goods to be valued. Secondly, biases and errors resulting from the use of hypothetic market in the CVM survey could be reduced substantially by reducing the bias of difference value of WTA and WTP. The use of dichotomous choice questionnaire(DCQ)in the survey will also be able to elicit more survey respondents’ true preference. It is suggested in this paper that we need to firstly promote various groups’ correct perception on the value of forest environment through the use of core valuation of differential designs, and secondly, to promote the research on typical forestry public goods and the integrated application of the improved and effective method of CVM.