Growing body of evidence suggests an association between circulating adiponectin and the development of atherosclerosis. However most of these studies have been performed with healthy subjects with relatively small sample size or patients with preexisting cardiovascular diseases. The results obtained are inconstant for an independent association between adiponectin and subclinical atherosclerosis. We aimed to evaluate the association between circulating adiponectin and carotid intima media thickness (CIMT) in a large sample of individuals from the community. We performed a non-diabetic community based cohort study conducted on a healthy Korean adults (age>40 years) population. The serum adiponectin concentrations were measured by a radioimmunoassay. The mean adiponectin level was significantly lower in subjects with high CIMT defined by >0.9 mm than those with low CIMT (9.18 vs. $10.44;{mu}g/mL$, P=0.03). The adiponectin level was correlated negatively with waist circumference (r=-0.27), body mass index (r=-0.21), systolic blood pressure (r=-0.11), diastolic blood pressure (r=-0.10), triglyceride (r=-0.30), LDL-cholesterol (r=-0.02), glucose (r=-0.16), insulin (r=-0.17), homeostasis assessment-insulin resistance (r=-0.23) and C-reactive protein (r=-0.11), and correlated positively with HDL cholesterol (r=0.30) (all P<0.05). Adiponectin was significantly associated with CIMT independently of age, sex, and all metabolic risk factors ($R^2$=0.14, P=0.04). In conclusion, serum adiponectin levels are independently associated with CIMT as a surrogate of subclinical atherosclerosis in a non-diabetic community based cohort.