國家公園的遊客一年達一千多萬人次，居戶外遊憩之主流地位，可見國家公園的遊憩效益之重要性。目前國內對國家公園遊憩效益的研究，除使用條件評估法之外，仍多採用單一地點評價的旅遊成本法；使用目前國外普遍使用的多點隨機效用模型的文獻，仍付之闕如。此外，目前國內文獻並無針對淡水魚以及昆蟲的物種多樣性之遊憩效益之估計，而此二者也是新興的生態旅遊之焦點。因此，本文將多點隨機效用模型，以及允許估計個人化參數之混合式羅吉特模型，應用在上述遊憩效益的估計上。本文的調查採面訪方式，估計結果發現這些遊憩效益相當高，且在各個國家公園之間，存在差異性。

Every year, more than ten million visitors visit the national parks of Taiwan. National park visiting is the main stream of outdoor recreation. Thus the recreational values of national parks are of importance. Literature about the recreational value of the national park has used the contingent valuation methods as well as single site travel cost methods. None of them use the currently popular multiple-site random utility models. In addition, none of them estimates the recreational values of fresh water fish species and insect species. These are focuses of evolving ecotourism recently. Thus this paper uses multiple-site random utility model and the mixed logit model, which is capable of estimating individual parameters, to estimate recreational these values. In-person survey is used. Estimation results demonstrate that the recreational values of national parks and the recreational values of animal species are heterogeneous across national parks, and in general, they are highly valued by tourists.