Background: We investigated the association between income-education groups and incident coronary heart disease (CHD) in a national prospective cohort study. Methods: The REasons for Geographic And Racial Differences in Stroke study recruited 30,239 black and white community-dwelling adults between 2003 and 2007 and collected participant-reported and in-home physiologic variables at baseline, with expert adjudicated CHD endpoints during follow-up. Mutually exclusive income-education groups were: low income (annual household income <$35,000)/low education (< high school), low income/high education, high income/low education, and high income/high education. Cox models estimated hazard ratios (HR) for incident CHD for each exposure group, examining differences by age group. Results: At baseline, 24,461 participants free of CHD experienced 809 incident CHD events through December 31, 2011 (median follow-up 6.0 years; interquartile range 4.5-7.3 years). Those with low income/low education had the highest incidence of CHD (10.1 [95 % CI 8.4-12.1]/1000 person-years). After full adjustment, those with low income/low education had higher risk of incident CHD (HR 1.42 [95 % CI: 1.14-1.76]) than those with high income/high education, but findings varied by age. Among those aged <65 years, compared with those reporting high income/high education, risk of incident CHD was significantly higher for those reporting low income/low education and low income/high education (adjusted HR 2.07 [95 % CI 1.42-3.01] and 1.69 [95 % CI 1.30-2.20], respectively). Those aged ≥65 years, risk of incident CHD was similar across income-education groups after full adjustment. Conclusion: For younger individuals, low income, regardless of education, was associated with higher risk of CHD, but not observed for ≥65 years. Findings suggest that for younger participants, education attainment may not overcome the disadvantage conferred by low income in terms of CHD risk, whereas among those ≥65 years, the independent effects of income and education are less pronounced.