Considering the strong correlation between carbohydrate and fat intake, we defined and assessed the association of the carbohydrate-to-fat ratio with the high-density lipoprotein cholesterol (HDL-c) level using 12-year follow-up data from the community-based cohort of the Korean Genome Epidemiology Study. We evaluated the long-term changes in HDL-c levels according to quartiles of carbohydrate-to-fat ratio using a mixed model. We also assessed the effect of the carbohydrate-to-fat ratio on the prevalence and incidence of hypo-HDL-cholesterolemia. Of 6,627 subjects, the prevalence of undiagnosed hypo-HDL-cholesterolemia at baseline was 35.3% (n = 2,339). Among the disease-free subjects, 56.8% developed hypo-HDL-cholesterolemia (incidence = 92/1,000 person-years). The prevalence and incidence of hypo-HDL-cholesterolemia were higher in females than in males. The highest carbohydrate-to-fat ratio quartile, which was characterized by high and low intake of carbohydrate and fat, was consistently associated with a lower HDL-c level during the 12-year follow up. Moreover, those in the highest quartile had a 1.14-fold greater risk of incident hypo-HDL-cholesterolemia than those in the lowest quartile, with a significant dose-response relationship. We found that high and low intake of carbohydrate and fat, respectively, was consistently associated with a low HDL-c level over a prolonged period. More research is needed to promote appropriate intake of macronutrients.