© 2019, Springer-Verlag GmbH Germany, part of Springer Nature. Aims/hypothesis: Ideal cardiovascular health (CVH) is associated with lower diabetes risk. However, it is unclear whether this association is similar across glycaemic levels (normal [<5.6 mmol/l] vs impaired fasting glucose [IFG] [5.6–6.9 mmol/l]). Methods: A secondary data analysis was performed in the REasons for Geographic and Racial Differences in Stroke (REGARDS) study. Incident diabetes was assessed among 7758 participants without diabetes at baseline (2003–2007) followed over 9.5 years. Baseline cholesterol, blood pressure, diet, smoking, physical activity and BMI were used to categorise participants based on the number (0–1, 2–3 and ≥4) of ideal CVH components. Risk ratios (RRs) were calculated using modified Poisson regression, adjusting for cardiovascular risk factors. Results: Among participants (mean age 63.0 [SD 8.4] years, 56% female, 73% white, 27% African-American), there were 891 incident diabetes cases. Participants with ≥4 vs 0–1 ideal CVH components with normal fasting glucose (n = 6004) had 80% lower risk (RR 0.20; 95% CI 0.10, 0.37), while participants with baseline IFG (n = 1754) had 13% lower risk (RR 0.87; 95% CI 0.58, 1.30) (p for interaction by baseline glucose status <0.0001). Additionally, the magnitude of the association of ideal CVH components with lower diabetes risk was stronger among white than African-American participants (p for interaction = 0.0338). Conclusions/interpretation: A higher number of ideal CVH components was associated with a dose-dependent lower risk of diabetes for participants with normal fasting glucose but not IFG. Tailored efforts that take into account observed differences by race and glycaemic level are needed for the primordial prevention of diabetes.