In Taiwan, the people who have joined National Health Insurance accounts for more than 99% of all. Therefore, the National health insurance database has become a typical and empirical data for related works. The National health insurance database has collected entire medical records of people. So, the records need be summarized appropriately before further data mining and related analysis. This thesis has presented a data warehouse system applying business intelligence for a case study for diabetes. A mesh-up method is used to explore the medical resources utilization for diabetes. In which method combines data mining, on-line analysis processing and multi-dimensional analysis from patients, medical institutions, and drugs.