There have been some discussions in China and elsewhere about the asymmetry between WTP and WTA values; however, so far no consistent conclusions have been reached. This is now a highly topical matter in relation to environmental impact assessment. Research on the subject is especially lacking in China, and is therefore, particularly necessary in this context. In view of this, a survey of 940 respondents was conducted in four cities of the Guangxi Beibu Gulf Economic Zone, in order to estimate individuals' maximum willingness to pay (WTP) for good coastal environmental conservation as well as their minimum willingness to accept (WTA) compensation through application of the logit mode. The purpose was to probe impact factors and mechanisms of difference between WTP and WTA, so as to render valuation methods more scientifically applicable and more reliable. The resultant improved methods will be suitable for use in the practice of environmental science in China. The results obtained showed the following. First, over five years, the expectancy values of WTP and WTA were 450.17 Yuan and 5772.71 Yuan per household per year, respectively, with WTA/WTP value of 12.8. This indicates obvious asymmetry but basically accords with empirical results of research in this field conducted both domestically and overseas. Second, through cross-analysis of WTP and WTA, it was found that in 29.4% of the sample (276 respondents), WTP and WTA were not both zero. The ratio of WTA to WTP was 12.54, with this very close to the estimation result of the logit model. This to some extent reflects differences of this group in estimating environmental products, while also proving the reliability of the logit model. Third, based on prospect theory analysis of behavioral economics, it was found that informants in the survey showed loss aversion and framing effects, and avoidance behavior of uncertain risk and expected ambiguity. At the same time, they also exhibited free-riding behavior. Finally, through analysis of situational factors, the group was divided into five grading levels (1-5), based on the five Richter scales used in the questionnaire. The average value of WTP was quite different to that of WTA. Of all situational factors, two had the greatest influence on respondents' choice of WTP and WTA, manifesting themselves through a punishing effect. The situational factor of economic income satisfaction had a significant impact on respondents' economic behavior, with the ratio of WTA to WTP showing the most obvious asymmetry in this case. The lowest and highest ratios both appeared with this situational factor. The ratio of WTA to WTP for those having the lowest level of satisfaction with their economic income was up to 32.5, far higher than the ratio for those who are basically satisfied and the ratio estimated from the logit model. This shows the loss aversion and punishing effects described in behavioral economics theory. However, the other three situational factors considered did not appear to exert influence to the same degree. In conclusion, the measurement of environmental value is complex, because it refers to hypothetical market valuation of public goods or quasi-public goods where there is no real actual market for reference; thus, people show more obvious irrational behavior. Local government should therefore pay good attention to local people's moods and attitudes when implementing environmental policy.