With the growing number of association rules, it becomes moreand more difficult for users to explore interesting rules due toits nature complexity. Studies base on human perception andintuition show that graphical representation could be a betterillustration of how to handle data by using the capabilities ofthe human visual system to seek information. The 3D matrixbasedapproach visualization system of association rules called3DMVS was implemented in present study. The main visualrepresentation employed the extended matrix-based approachwith rule-to-items mapping to general transaction data set. Anovel method merging rules and assigning weight is proposedto generate new rules to reduce the dimension of theassociation rules, which will help users to find more importantitems in the new rule. Additionally, several interactions suchas sorting, filtering, zoom and rotation, facilitate decisionmakers to explore the rules they are interested in variousaspects. Finally, various evaluation techniques have beenemployed to assess the system from a logical reasoning pointof view.