Objectives To examine the association between Mini-Mental State Examination (MMSE) score and motor vehicle crash (MVC) risk in a large cohort of community-dwelling participants with cardiovascular disease (CVD) or diabetes mellitus. Design Prospective observational study. Setting Participants enrolled in the Ongoing Telmisartan Alone and in Combination With Ramipril Global End Point Trial and Telmisartan Randomized Assessment Study in ACE Intolerant Subjects with Cardiovascular Disease clinical trial, which included individuals aged 55 and older with CVD or diabetes mellitus. Participants Totally 17,538 frequent drivers (defined as driving at least once per week) who had completed a baseline MMSE. Measurements Involvement in a MVC as the driver. Results Baseline MMSE score was divided into four categories: 30, 27-29, 24-26, and < 24. The median MMSE score was 29 (interquartile range 27-30), and 726 (4.1%) has a MMSE score of less than 24 at baseline. After a mean follow-up of 4.5 years, 1,068 (6.1%) participants were drivers in a MVC. Lower scores were not associated with future MVCs (MMSE score 29-27, hazard ratio (HR) = 1.06, 95% confidence interval (CI) = 0.93-1.22); MMSE score 26-24, HR = 0.96, 95% CI = 0.78-1.19; MMSE score <24, HR = 0.72, 95% CI = 0.50-1.05) on multivariable analysis. A MVC within the previous 2 years (HR = 2.68, 95% CI = 2.29-3.13) was the strongest predictor of future MVCs. Other clinical factors associated with greater MVC risk were depression, falls within the previous year, sleep apnea, and lower baseline systolic blood pressure. Conclusion In a population of frequent drivers, the MMSE does not predict MVCs. Other clinical factors have a stronger association with MVC risk.