Abstract: Discovering association rules and making prediction with data mining methods are two hot\|spots in the domain of dealing with information. In this paper, the normal cloud model is adopted to replace the domain partition of quantitative attributes and the concept of normal cloud rules on the quantitative attributes is given. Then the method of mining the normal cloud association rules is provided and the mined rules are used for prediction. Because the normal cloud model can well soften the domain partition boundary of the quantitative attributes, the normal cloud association rules and the prediction results can be easily understood.