在當代的各種科技風險之中，核能發電是極具爭議且備受注目的議題，尤其當日本福島核災發生之後，社會對核電之「低發生機率、高危害結果」的風險特質愈加敏感。在此種緊張的社會氛圍下，找出影響民眾核電政策偏好的關鍵因素，成為能否有效處理核電爭議的重要前提。本文認為針對這種風險特質較為特殊的政策，如欲有效地辨識民眾偏好的成因，不應侷限於傳統的理論框架。為填補目前核能風險研究的缺口，本研究結合過去科技風險與環境主義的文獻，嘗試提出一個整合性的研究架構，並透過實際資料的佐證，藉此釐清民眾之風險感知的來源及政策偏好的成因，以作為未來政府進行風險溝通之參考。研究結果顯示，本研究所提出的整合性架構，對民眾核電偏好的解釋力為47.5%，一方面證實民眾的議題知識、機構信任、電價承受、環境信念及風險感知對於其核電政策偏好有顯著影響；另一方面，則是確認風險感知對於機構信任、電價承受及環境信念的中介效果。以上成果均是過去核能風險研究鮮少提及與驗證的部分。在此一基礎之上，本研究認為政府惟有改變過去風險溝通與管理的思維，學習從各類利害相關人的觀點去分析與理解風險議題，才能確實掌握問題的癥結與對策；同時，政府更應設計與開放各種制度性溝通與參與的管道，讓不同的利害相關人能夠充分地參與核能發電相關政策的討論與決策過程，才有助於化解長久以來的能源爭議。 Nuclear energy is one of the most salient and controversial technological risks in modern society. The reasons for this stem from its characteristics of low-probability and high-consequence. Recent disasters, such as the one in Fukushima, Japan, have also further exacerbated citizen concerns over nuclear technology. In light of these developments, the understanding of the reasons behind public preferences toward nuclear energy is a necessary step toward implementing energy policies that can best approximate citizen expectations and national needs. In order to bridge past research deficiencies, this research has combined the point of view of technology risk and environmentalism in the construction of an integrative framework. Empirical data on public opinion were utilized to identify the variables behind citizens’ nuclear energy policy preferences. The results may serve as references for the implementation of Taiwan’s nuclear energy risk communication. The adjusted coefficient of determination of the model is 47.5%. The results revealed that citizens’ knowledge on policy issues, trust in government, price tolerance of electricity, environmental belief in nuclear energy and risk perception all have significant influences on the formulation of policy preferences. Risk perception was proven to function as an immediate variable whose influence allows trust in the government, price tolerance of electricity and environmental belief in nuclear energy, to have direct and indirect effects on policy preference. These findings have rarely been discussed or verified by past studies. This research suggests that the government should restructure its approaches to risk management and communication, thereby learning to analyze and comprehend the crux of the problem through the point of view of stakeholders. The government should also design and open up different types of communication and participation channels, allowing stakeholders to engage in the discussions and decision processes of nuclear energy policies. Communication should not be limited to the interagency or interparty levels. Only when the dual problems of approach restructuring and stakeholder communication are addressed will tangible improvements be made.