Congestion is a type of eternality that causes marginal social benefit (MSB) to differ from marginal social cost (MSC). For efficient resource allocation, the congestion effect should be measured and accounted for in the decision making procedure. Congestion in a forest-based recreational site seems more important not only because of rapid growth in site demand, but also because the recreational site is not producible by man. Further, forest-based recreational sites usually have been provided by the public sector at nominal or zero prices, independent of supply and demand for the resources. Therefore, no price mechanism exists to control effectively the supply of and demand for recreational site. When congestion exists in a recreational site, the duality of the recreational experience obtained by users deteriorates causing reduced willingness to pay by users. The sum of reduction of individual willingness to pay can be thought of as a marginal social cost associated with a specific level of congestion for a specific recreational site (Cicchetti and Smith 1976). In this context, the purposes of this research are (1) to estimate a willingness to pay function for an individual using a high-density forest-based recreational site and (2) to apply the willingness to pay function to aid in the management of the Book Han San National Park in Korea. The change in individual willingness to pay in response to different level of congestion in a recreational site can be explained by the theory of club as well as conventional modeling approaches for consumer behavior. That is, the congestion effect can be captured by the individual demand for the site visit derived by maximizing his utility function subject to budget constraint (McConnell, 1977). To estimate the congestion effect, however, we cannot use market-oriented data because congestion is not a market good. There are two kinds of methods available to evaluate a non-market good ; the direct and indirect methods. The indirect method uses the behavior of consumers in related markets to estimate the users` valuation of non-market good. The direct method is to ask individuals to reveal his or her willingness to pay for the hypothetical situation of varying degrees of congestion for the recreational site (Just et al, 1982). Even though both methods generally can he used to evaluate the site qualityb variable, the direct method seems to be the only practical method in some situations and that is the case for the empirical study (McConnell, 1977: Cicchetti and Smith, 1976). To be able to ask an individual, using direct method, to state the change in his willingness to pay, it is necessary to describe the congestion situation objectively and perceptibly. None of the descriptions of congestion (or measure of congestion) used in previous studies seem to be appropriate for use in a high-density recreational site. To solve this problem, a visual aids, specifically using video taped segment, were developed, in which five different levels of trail congestion were described in terms of perceived user density and noise level on two different trails of oaring route. Using ordinary least square (OLS) and generalized least square (GLS) methods, individual willingness to pay was regressed on various independent variables representing site quality, recreational behabior, attitude toward natural environment and socio-economic status with linear and semi-log specifications. Among four equations estimated and reported in Chapter IV, OLS semilog specification was used for the further discussion of an application of the estimated willingness to pay relationship. The final willingness to pay equation estimated with OLS semi-log specification is : In (WTP) -7.903-0.273 UD+0.161 ENVR₁+ 0.295 ENVR₂+0.036 TT-0.058 ED+0.069 FS-0.016 DA (R²=0.15) (6-1) Where the variables included are defined in Table IV-2. Then, the willingness to pay function was used to determine the optimal number of users for the