It has been demonstrated that higher degree of arteriolar retinopathy is associated with greater cardiovascular risk, and hyperhomocysteinaemia is also related to increased cardiovascular risk, but interacts with other risk factors, particularly smoking. It still remains unclear regarding relationships of smoking, fasting plasma total homocysteine (tHcy) levels and arteriolar retinopathy. This study was aimed to investigate the relationship and influence of smoking and tHcy levels on degree of arteriolar retinopathy. Two hundred and forty-three subjects were enrolled from an annual health examination. The arteriolar retinopathy was examined by direct ophthalmoscopy. Dundett ANOVA showed that geometric mean of tHcy levels were 11.5+/-1.54 versus 11.2+/-1.41 versus 17.6+/-1.92 (P1=0.883, P2=0.001) in subjects with no arteriolar retinopathy (as control group), grades I and II retinopathy, respectively. Furthermore, multiple linear regression analysis showed that only smoking consumption (P<0.001), gender (P=0.012) and presence of hypertension (P=0.041) were independent determinants of plasma tHcy levels. After females were excluded, T-test showed a significant differences in tHcy levels (15.6+/-1.56 micromol/L versus 12.4+/-1.45 micromol/L, P=0.003) and in prevalence of grade II retinopathy (25.4% versus 9.3%, P=0.029), but no difference in other variables or prevalence of overall retinopathy between smokers and non-smokers. Finally, logistic regression showed that smoking (OR 4.19, 95% CI 1.17-15.0) was a stronger predictor than hyperhomocysteinaemia (OR 2.14, 95% CI 0.85-5.41) for presence of grade II retinopathy. This study showed that smoking was related to increased plasma tHcy levels in subjects with grade II retinopathy, and it could independently contribute to facilitating the progression of arteriolar retinopathy.