Background: It is not clear whether offering health checks to the general population can be used to prevent diabetes. Few randomized studies have had a long-term follow-up. We used a randomly selected population cohort as a randomized trial to examine the effect of repeated health checks on the 30-year incidence of diabetes. Methods: The study included all persons from 11 municipalities in Copenhagen aged 30, 40, 50, and 60 years (n = 17845). An age-stratified and gender-stratified random sample (N = 4789) was invited to participate in a maximum of three health checks between 1982 and 1994 (‘intervention group’). The remaining 12994 persons were defined as the ‘control group’. The health checks included a questionnaire, a physical examination including assessment of overweight and blood pressure, and blood sampling with determination of serum lipid levels. Based on the person’s answers and test results, the participants were given individual information about the results, disease risk and lifestyle. Their general practitioner, too, was provided with written information on the test results. Both groups were followed in the Danish Civil Registration System, the Danish National Diabetes Register, the Cause of Death Registry, and the National Patient Registry until 31 December 2012. Findings: There were 2636 incident cases of diabetes and a mean follow-up time of 24.1 years. The age-adjusted and gender-adjusted hazard ratio (HR) (95% confidence interval, CI) for the intervention group versus the control group was HR = 1.07 (95% CI: 0.98, 1.16, p = 0.153). Interpretation: Offering repeated general health checks to the general population had no preventive effect on the development of diabetes during 30 years of follow-up.