Objective To explore vitamin D status in relation to diabetes, based on data from a national health examination survey in Thailand. Design and Methods A total of 2,641 adults, aged 15–98 years, were randomly selected according to geographical region from the Thai 4th National Health Examination Survey sample. Logistic regressions were used to examine the cross-sectional association between diabetes status and level of 25(OH)D separately by age groups and areas area of residence. Results Fifty percent of the subjects were male and 5.8% had diabetes. The mean level of 25(OH)D was 79.3 ± 0.8 nmol/L. Six percent and 45% had vitamin D insufficiency based on cut-off values of 50 and 75 nmol/L, respectively. In a regression model, it was found that 25(OH)D3 and total 25(OH)D were positively associated with diabetes. In addition, logistic regression analysis showed low circulating 25(OH)D3 but not 25(OH)D2 levels was significantly associated with an increased odds of diabetes in older persons (aged >=70 years) in urban areas. However, for subjects residing in rural areas, no association between serum 25(OH)D3 or total 25(OH)D and diabetes was found. Furthermore, vitamin D insufficiency was associated with a higher risk of diabetes (OR 1.56, 95%CI 1.10, 1.12) only in the urban elderlys Conclusion Low vitamin D status is modestly associated with a small increase in the risk of diabetes in urban Thai elderly. The observation that higher vitamin D status is associated with increased diabetic risk in young adults needs to be further explored and confirmed.