PURPOSE: To examine the association between serum homocysteine and retinal emboli in a general older population. DESIGN: Cross-sectional population-based study. METHODS: Homocysteine level was determined from fasting venous blood collection at the second Blue Mountains Eye Study (BMES) (n = 3509, age > or =49 years, 1997 to 2000). Retinal emboli were graded from retinal photographs. Using logistic regression, we analyzed associations between serum homocysteine and retinal emboli adjusting for age, gender, hypertension, body mass index, and lipid levels. RESULTS: After multivariable adjustment, serum homocysteine was associated with presence of retinal emboli (odds ratio (OR) 1.2; 95% confidence interval (CI), 1.0 to 1.6 per standard deviation [4.8 micromol/l] increase in serum homocysteine). CONCLUSION: Elevated serum homocysteine is weakly associated with increased odds of retinal emboli in this older population-based study.