The contingent valuation method (CVM) is a survey-based method that uses random samples of families or individuals to reveal consumer preferences for goods or services, such as environmental resources, in a hypothetical market. Through questionnaires, CVM identifies people's preferences and expresses them in dollar terms as willingness to pay (WTP) or willingness to accept (WTA). WTP (or WTA) can be used to gain insights into the potential economic benefits or losses brought about by a planned project. A great variety of questionnaire formats have been developed and applied, among which the dichotomous choice CVM is considered to be one of the most advanced methods. In recent years, the double-bounded dichotomous choice contingent valuation method (DBDC-CVM) has been widely used to value the ecological environment and natural resources. It can identify respondents' true WTP and reduce the deviation of assessment results, thereby making the results more accurate and reliable. The initial bid number and sample size are very important determinants of WTP in DBDC-CVM. Fewer initial bids and smaller sample sizes will reduce the validity and reliability of the resultant WTP. In contrast, more initial bids and larger sample sizes increases the difficulty and cost of the survey. The number of initial bids is always obtained using a pre-test, while the sample size is always determined by field research or online surveys. This study investigates the dynamic effects of initial bid number and sample size on mean WTP, using the willingness to pay (WTP) function model and a-posteriori Monte Carlo simulation. We use Weibull and log-logistic distributions of WTP to simulate two different sources of data. In order to simplify the Monte Carlo simulation, we quantize the probability distribution and make a hypothesis to design bids. We produce four graphs to visualize the simulated results and use these to identify the optimum number of initial bids and sample size. Initial bid number and sample size had almost the same influence on mean WTP for both WTP distributions. When the initial bid number t and sample size exceed 5 and 500, respectively, the influence on mean WTP is almost the same. Based on this result we suggest that at least 5 initial bids and 500 samples are needed in order to ensure the validity and reliability of estimates of mean of WTP generated by Monte Carlo simulation. These results provide a quantitative basis for CVM questionnaire design and have the potential to improve the validity and reliability of WTP.