We present an adaptive conversational agent system that considers user preferences to make rapport with a user through a conversation. Rapport is a phenomenon that we can feel the connection and comfort through a conversation, defined in social psychology. Satisfactory interaction, talking about the intimate topic for participants, eye gazing, and attentively listening are effective measures to build rapport. In this presentation, we focus on this satisfactory interaction and propose an adaptive conversational agent that talks satisfactory with considering user preference. If user preferences for responses of conversational agent systems are fulfilled, the system response evokes high engagement of users to continue a conversation in the long term, and high satisfaction to make the conversation more comfortable. The types of preference depend on users, and types of user preference have a variety: speaking style, dialogue strategies, and communication distance. To consider these many different user preferences, the method that handles each type of user preference appropriately is necessary. To solve this problem, we tried four cooperative factors for example-based dialogue modeling. First, we proposed a linguistic individuality transformation method to transform the speaking style of conversational agent’s responses as the user hopes. This method makes it possible for the conversational agent to talk with the individuality that the user hopes. Second, we propose a satisfaction pr diction method for example database that the conversational agent holds inside, to achieve talks with high satisfaction. This method makes it possible to select a response of conversational agent that increases the user satisfaction. Third, we propose an adaptive response selection method with considering user feedbacks to select the best response for the user preference. This method achieves that the conversational agent responds with the best satisfactory response for the user. Fourth, we propose a response selection method based on entrainment analysis to select an appropriate response on entrainment depends on dialogue act. Entrainment is a phenomenon which called synchrony, relates to the naturalness of conversation. This response selection method based on entrainment analysis makes it possible for the conversational agent to select an appropriate response with considering the entrainment given dialogue act. We examine these cooperative factors and indicate the effectiveness of proposed methods from each experimental evaluation.