We examine an intelligent supervisory level control by using a qualitative, physical model of a steam generator in this paper. And the effectiveness is shown by a HIL simulation in the meaning of included a target value adjustment of an expert in a control loop. It is shown that the three elements control has an effect of suppressing pressure changes. If the gain is raised too much, it becomes unstable. However, this instability might be controlled by a target value operation by an expert method. Moreover, the steady state deviation in the case of more high target value can be suppressed by a 4-elements control with a new feedback loop of generation calorie. Then, a basic specification of a simple fuzzy target valued control is shown.