Congestion is a type of externality which causes marginal social cost to differ from marginal social benefit. To achive Pareto efficient resource allocation, the congestion effect should be measured and accounted for in the decision making procedure;Using the direct interview method, the effect of congestion in a high-density forest-based recreational site in Korea was measured in terms of the change in individual willingness to pay for varying levels of congestion and social demographic characteristics. Data were collected using a questionnaire from a random sample of about 221 recreationists during Fall, 1987;Video tape segments illustrating five different levels of congestion were used by the interviewers to elicit responses on the willingness to pay;Individual willingness to pay was estimated by the ordinary least square (OLS) and generalized least square (GLS) estimation method;The OLS semi-log specification is: ln(WTN) = 7.903 - 0.273 UD + 0.161 ENVR1 + 0.295 ENVR2 + 0.036 TT - 0.058 ED + 0.069 FS - 0.016 DA, where WTP = individual willingness to pay, ENVR1, ENVR2 = factor scores measuring attitude toward natural environment, TT = travel time to reach the site, ED = education, FM = family size, and DA = days available for recreational site use;This willingness to pay relationship was applied to determine the optimal number of users for the particular forest-based recreational site;The optimal use level is calculated as about 73,000 users per day. Then, the managerial implication of the optimal user level was discussed.