Objective: This study aimed to identify the best predictor of metabolic syndrome (MetS) by comparing the predicting ability of various anthropometric and atherogenic parameters in the Chinese Han population in Xinjiang. Methods: A representative, cross-sectional sample of 5,757 Chinese Han adults were selected from the Cardiovascular Risk Survey conducted from October 2007 to March 2010. MetS prevalence, area under the curve (AUC), distance on the receiver operating characteristic curve and the cut-offs of each variable were compared for the presence of MetS. Results: According to the criteria of the International Diabetes Federation (IDF), the Joint Interim Statement (JIS) and the Third Adult Treatment Panel (ATPIII), 32.1, 48.5, 39.3% of men and 38.0, 45.1, 44.9% of women had MetS in Xinjiang. According to the IDF criteria, the waist-to-height ratio (WHtR) had the highest AUC value in men (0.836) and women (0.837), with the optimal cut-off of 0.54 in men and 0.53 in women. According to both the JIS and ATPIII criteria, triglycerides/high-density lipoprotein cholesterol TG/HDL-C had the highest AUC value in men (0.830 and 0.833, respectively) and women (0.832 and 0.827, respectively), with the optimal cut-offs being 1.6 and 1.2 in men and 1.1 and 1.1 in women, respectively. Conclusion: WHtR was the best predictor of MetS according to the IDF criteria while TG/HDL-C was the best predictor of MetS according to the JIS and ATPIII criteria.