Apart from the traditional recommendation system which orients to computer, personalized recommendation system (PRS) adapts to user. Thus, there is a need to setup a theoretical system of user-centered personalized recommendation system in accordance with real applications. This paper constructs the system with 5 parts: disciplinary basis, stage of development, essence, architecture and assessment. Specifically, the paper proposes that the essence of personalized recommendation system is cognitive assistant and the H3W system architecture wi...