The purpose of this study is to propose a new method: decision model based on grey system theory and rough Sets (DGR). This method can improve the shortcomings of the subjective judgment of the decision attribute in the rough sets by the calculations of reachable grey gtructural modeling (RGSM). The result shows that DGR can not only determine the influential objects from different dividing clusters through RGSM, but also find the core of attributes in the information table by rough sets. DGR can be applied to analyze data and decision-making problems. It not only to be applied in decision-making, but also can in the field of data mining and knowledge discovery.