Metabolic syndrome (MetS) is a major public health concern. Efficient screening requires criteria that are economical, easily accessible, and applicable for all populations. We aimed to compare the discriminating ability of body mass index (BMI), waist circumference (WC), and waist-to-height ratio (WHtR) for the diagnosis of MetS in a Han Chinese population. Demographic information, obesity indices, and results of biochemical tests were collected from a cross-sectional sample of 8084 individuals (3619 men and 4465 women, 18-79 years old) from Changping District, Beijing, China. Areas under receiver operating characteristic curves (AUCs) and adjusted odd ratios of 3 obesity indices were analyzed and their optimal cutoffs were determined. For women, the AUCs demonstrated that WHtR was significantly more powerful than BMI and WC (both P<.05) for predicting MetS [WHtR, 0.857 (0.846-0.868); WC, 0.849 (0.837-0.860); BMI, 0.808 (0.795-0.821)]. For men, WHtR was significantly better than BMI [P<.05; WHtR, 0.859 (0.846-0.871); WC, 0.855 (0.843-0.868); BMI, 0.815 (0.802-0.829)]. The optimal cutoffs for WHtR for discriminating MetS were 0.51 in both genders. Multiple logistic regression confirmed the positive association between WHtR and the risk of MetS. In the nonobese subgroup, WHtR was also superior to BMI and WC for predicting MetS in men (P<.05) and better than BMI in women (P<.05). Among the obesity indices analyzed here, WHtR was the best for predicting MetS in Han Chinese adults, especially in nonobese adults.