The Internet security has come to be regarded as a critical concern of the survival and success of the burgeoning businesses on the net [1]. Although many security solutions have been introduced to deal with some specific threats, an automated and comprehensive protecting mechanism is still in need. To achieve this end, artificial intelligent techniques are believed to be applicable and promising. Accordingly, a multi-agent Internet security system was proposed [2,3]. It is the goal of this paper to study the means of constructing a formal conceptual model of generic Internet security processes, which would help to understand and analyze the knowledge in the domain of the Internet security, and to facilitate the building of the proposed multi-agent security system.