In order to implement a decision-making mechanism in an intelligent system, a logical reasoning structure is usually supposed to play the major role. So far, this fashion was successful in the sense that it fits the computer programming. But for a human-oriented man-machine interface, it is required to incorporate intuitive human ways in determining what action to take for adjustment, and it leads to more flexible responses. Here in this paper, discussed are knowledge and rule structures together with the concept of context as basic stuctures to deal with learning and reasoning behind decision-making.