Introduction: There are different methods to assess the fat body excess, but whether one of them is better to detect type 2 diabetes mellitus (T2DM) cases has not been fully explored in Peru. This study aimed to explore the diagnostic accuracy of some obesity anthropometric markers for newly-diagnosed T2DM at the population level and by sex. Methodology: Secondary data analysis conducted using data from a population-based study carried out in Tumbes, Peru, with subjects aged between 30 and 69 years old. The outcome was newly diagnosed T2DM, defined using an oral glucose tolerance test. The index tests were obesity anthropometric markers: body mass index (BMI), waist circumference (WC), and waist-to-height ratio (WHR). Diagnostic accuracy of anthropometric markers was estimated using the area under the ROC curve (AUC); sensitivity and specificity were also estimated based on the Youden index. Results: Data from 1,500 participants were analyzed; and 4.7% were classified as having undiagnosed T2DM. Mean age was 47.6 (standard deviation: 10.6) and 50.1% were male. At the population level, diagnostic accuracy of anthropometric markers was: WHR (AUC: 0.67; 95%CI: 0.60-0.73), BMI (AUC: 0.65; 95%CI: 0.58-0.72), and WC (AUC: 0.65; 95%CI: 0.58-0.72). Stratified by sex, the results were similar except on the case of male WC, with an acceptable diagnostic accuracy (AUC: 0.70; 95%CI: 0.60-0.81). Conclusions: The diagnostic accuracy of obesity anthropometric markers (BMI, WC and WHR) for T2DM screening was poor in the general population