Public acceptance has become an important factor in nuclear power program particularly after Chernobyl accident and recent rapid democratization in Korea. Methods reflecting public opinions in order to improve public acceptance are firstly to understand what the public think about nuclear power plant and secondly to find out the public preference values for its policies. For this purpose, simplified multi-attribute utility (MAU) model was applied to analyze the public perception pattern for fire power production systems. And the conjoint analysis was applied to find out the quantitative values of public preferences for twelve policy alternatives to improve the safety and to support communities surrounding nuclear power plants in Korea. To implement these perception and preference analyses, mail survey was conducted to the Qualified sample who had the experience of visiting nuclear power plant. Diagnosis of their perception pattern for five power production systems was made by the simplified MAU model. Estimation of the quantitative preference values for potential policy alternatives was made by the conjoint measurement technique, which made it possible to forecast the effectiveness of each option. The results from the qualified sample and the methods used in this study would be helpful to set up new policy of nuclear power plant.