To check and analyze the effectiveness of knowledge base system is a very important issue for implementing the promises of AI technology. This issue has been largely ignored by researchers in this field. This paper will focus on and present the consistency problem, which is a important aspect of knowledge base effectiveness analysis, of knowledge base systems. The 'classical' definition for consistency of a formal system is inadequate for many cases under knowledge base system context. The earlier work by others have roughly defined the consistency problem of knowledge base systems, and given some simplified results. We will develop a qualitative model for the knowledge base systems, and investigate some properties of the rule set of the knowledge base. This model can be used for specification and analysis of the behaviors of the knowledge base systems. Based on this model, we will further identify and clarify the consistency problem of knowledge base system and propose a model-based methodology used for checking and analyzing the consistency of a given knowledge base system. This model-based checking methodology may help for the design and implementation of the knowledge base systems in many applications.