This study aimed to clarify whether body mass index (BMI), waist/hip ratio (WHR) or percent body fat (%fat) is associated with coronary artery disease (CAD) in Chinese type 2 diabetic patients in Taiwan. A total of 463 patients were recruited. BMI and WHR were measured by standard methods and %fat by bioelectrical impedance. CAD was diagnosed as acute myocardial infarction, angina pectoris, or an electrocardiogram showing 'coronary probable or possible' according to the Minnesota codes. Age, sex, diabetes duration, hypertension, smoking, fasting plasma glucose, hemoglobin A(1c),and serum concentrations of total cholesterol, triglyceride, high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol were treated as confounders. Results showed that 144 patients had CAD. Age, hypertension, HDL-C and %fat were independently associated with CAD. CAD prevalence was 25.5%, 26.8%, 31.9% and 43.0%, respectively, for the first to fourth quartile of %fat (p<0.05). Multivariate-adjusted odds ratio for CAD for every 1% increase in %fat was 1.02 (1.01-1.03); and 1.01 (0.73-1.88), 1.26 (0.69-2.32) and 2.11 (1.09-4.07) for the second to fourth quartile, respectively, compared with the first quartile. BMI and WHR were not associated with CAD in similar analyses. In conclusion, %fat was a better predictor for CAD than BMI and WHR in Chinese type 2 diabetic patients in Taiwan.