Objectives: In the structural reform bill of health care, which passed the Diet in fiscal year 2006, the number of patients with lifestyle-related diseases and the number of those who will potentially develop such diseases in 2015 should be reduced by 25% from the number in 2008 through the national intervention program against obesity. We estimated the reduction in prevalence of diabetes mellitus, as a representative lifestyle-related disease, and the health care costs reduced by controlling obesity. Methods: Firstly, we estimated the prevalence (95% confidence interval) of obese people by conducting the National Health and Nutrition Examination Survey in 2005. Secondly, we estimated the proportion of obese people that should be reduced in order to reduce diabetes prevalence by 25% using the data from the National Diabetic Patients Survey in 2002. Thirdly, we estimated changes in prevalence of diabetes mellitus when the proportion of obese people was reduced by 20%, 40%, 60%, 80%, and 100%. Finally, we estimated how much health care costs would be reduced if the number of obese people was reduced by 20%. Results: It is extremely difficult to reduce the prevalence of diabetes mellitus by 25% by only reducing the proportion of obese people. From our estimation of changes in the prevalence of diabetes mellitus when the proportion of obese people was reduced, the intervention for people aged from 40 years to 59 years was more effective than that for people in other age groups for both male and female. The health care costs of diabetes mellitus can be reduced by ¥841,210,000,000 for male and by ¥75,930,000,000 for female. Conclusion: It is almost impossible to reduce the prevalence of diabetes mellitus by 25% although it is cost-effective to target on people aged from 40 to 59 years against obesity to reduce the prevalence of diabetes mellitus.