In this paper, the method how to configure the proposed new CAI system, which has been under development, was described. This system has aimed at the effective education of trainees to master knowledge and skill for diagnosing nuclear power plant (NPP) anomalies. In this study, the authors aim at configuring this system based on virtual reality technology and bio-informatic sensing methodology, in order to promote intuitive learning of user as well as to endow the flexible adaptation to the user¡Çs mental and physical state. There are two special features for realizing the target CAI system. (i) The utilization of Head-Mounted Display (HMD) provides the use with high training environment, by presenting 3D images of plant layout, and (ii) Application of various bio-informatic sensing techniques would be effective to inter various human internal states. Elemental studies needed for realizing the adaptation mechanism of this CAI system were described in this paper.