The accuracy of the Framingham risk score (FRS) in identifying patients with nonalcoholic fatty liver disease (NAFLD) at higher 10-year coronary heart disease (CHD) risk remains unknown. We aimed at evaluating both the baseline probability of CHD as predicted by the FRS and the actual long-term occurrence of CHD in NAFLD patients. This was a longitudinal study of a community-based cohort. A total of 309 NAFLD patients were followed up for 11.5 ± 4.1 years (total 3554 person-years). The overall calculated 10-year CHD risk was significantly higher in the NAFLD cohort than the absolute CHD risk predicted by the FRS for persons of the same age and gender (10.9 ± 9.3% vs. 9.9 ± 5.9%, respectively, P < 0.0001), and higher in men than women (12.6 ± 10.3% vs. 9.6 ± 8.1%, respectively, P = 0.006). New onset CHD occurred in 34 patients (11% vs. 10.9% predicted at baseline, P = NS), whereas 279 (89%) patients did not develop CHD. Using multivariable analysis, the FRS was the only variable significantly associated with new onset CHD (OR = 1.13, 95% CI = 1.05–1.21; P = 0.001). A FRS cut-point of 11 in women, and 6 in men had a sensitivity of 80% and 74%, respectively, and a negative predictive value of 97% and 93% respectively. NAFLD patients have a higher 10-year CHD risk than the general population of the same age and gender. The FRS accurately predicts the higher 10-year CHD risk in NAFLD patients, and helps identify those patients expected to derive the most benefit from early intervention to prevent CHD events.