The objective of the study was to assess the prospective association between serum uric acid levels and incident nonalcoholic fatty liver disease in a cohort of healthy Korean men. A cohort study was performed on 5741 Korean men, 30 to 59 years of age, with no evidence of fatty liver disease on liver ultrasound and with no major risk factors for liver disease at baseline. Study participants were followed in annual or biennial health examinations between 2002 and 2008. The presence of fatty liver was determined at each examination by ultrasound. Cox proportional hazards models were used to evaluate the association of baseline and time-dependent levels of serum uric acid with incident fatty liver, adjusted for potential confounders. During 23,995 person-years of follow-up, 1717 participants developed fatty liver on ultrasound examination. After adjustment for age, body mass index, smoking, and alcohol intake, the hazard ratios (95% confidence intervals) for incident fatty liver comparing quartiles 2 to 4 of serum uric acid to quartile 1 were 1.17 (1.01-1.37), 1.28 (1.11-1.48), and 1.51 (1.31-1.73), respectively (P for trend = .001). The adjusted hazard ratio comparing participants with hyperuricemia (serum uric acid ≥7.0 mg/dL) to those with normouricemia (<7.0 mg/dL) was 1.29 (1.14-1.46). A graded and statistically significant association persisted after adjusting for other cardiometabolic factors and also in time-dependent models. Serum uric acid was an independent risk factor of incident fatty liver detected by ultrasonography. Additional research should clarify the mechanisms underlying this association and the role of hyperuricemia in the development of fatty liver.