The invention discloses a real-time application method for multi-level storage data. The method comprises the steps of 1, service object modeling; 2, distributed object caching; 3, distributed data demand activation; 4, dynamic routing; 5, agent-based computing; 6, summarizing and aggregating calculation. Compared with the prior art, the method has the advantages that extend property is high, the real-time performance of tasks is high, multi-dimension multi-angle data analysis and inquiry are flexible, real-time data extraction is provided for OMS system statement presentation, manual dispersed data reporting is avoided, statement dimensionality and inquiry conditions can be changed quickly, and drag-and-drop establishment of an inquiry statement can be achieved in a user-friendly mode.