The objective was to examine the association between lifestyle and risk for diabetes. For an average of 9.9 years, this study prospectively followed a cohort of 7,211 (2,524 men and 4,687 women) community residents aged 30–69 years without diabetes at a health check-up conducted between April 1990 and March 1992 until diabetes was confirmed or until the end of 2006. The subjects were divided into 6 groups according to their total scores of Breslow’s lifestyle index (1–2, 3, 4, 5, 6 and 7 points). The association between lifestyle and diabetes incidence was investigated using Cox proportional hazards regression models. The results showed that the multivariate-adjusted hazard ratios were 0.45 in subjects who scored 5 points, 0.39 in subjects who scored 6 points, and 0.31 in subjects who scored 7 points, compared with subjects who scored 1–2 points. These data indicate that the healthy behaviors prevent the incidence of diabetes.