Data Mining is an emerging technology that intelligently assists users analyzing vast amounts of data and finding interesting nuggets of useful knowledge. Since the Data Mining process is by definition interactive and iterative, it is a challenge to provide a high-performance, rapid-response environment that assists users in all steps of Data Mining process to achieve their goals. Therefore, it is necessary to build an integrated Data Mining environment that can accept domain knowledge provided by users and utilize it as backgroud knowledge in the proper selection of input-output variables and matching of appropriate tools and techniques. In this regards, this thesis presents a modeling methodology using Qualitative Reasoning theory that models the underlying causal structures and commonsense in the target domain. The resulting Qualitative Model is used as background knowledge in Data Mining process. By adopting Qualitative Model, a complex system in real world can be reduced in a simple representation by focusing in its structure based on causal relationship. Reasoning with Qualitative Models propagates changes through the qualitative calculus upon causal relationship among their components. Thus, Qualitative Reasoning is an intuitive and natural way of reasoning as the human does. Stock-QM is developed as an example system that uses Qualitative Model as background knowledge in the Data Mining process. Stock-QM has domain knowledge about the trading strategy for the composite price index of Korean stock market-KOSPI. Stock-QM enables users to qualitatively test their assumptions on the move of KOSPI and to quickly identify what are influential causal factors. The functions of Stock-QM is a good example that shows how to mitigate the iterative nature inherent in Data Mining process.