Nonalcoholic fatty liver disease (NAFLD) is the most prevalent form of chronic liver disease in the world. New non-invasive diagnostic tools are needed to promptly treat this disease and avoid its complications. This study aimed to find key metabolites and related variables that could be used to predict and diagnose NAFLD. Ninety-eight subjects with NAFLD and 45 controls from the Fatty Liver in Obesity (FLiO) Study (NCT03183193) were analyzed. NAFLD was diagnosed and graded by ultrasound and classified into two groups: 0 (controls) and ≥ 1 (NAFLD). Hepatic status was additionally assessed through magnetic resonance imaging (MRI), elastography, and determination of transaminases. Anthropometry, body composition (DXA), biochemical parameters, and lifestyle factors were evaluated as well. Non-targeted metabolomics of serum was performed with high-performance liquid chromatography coupled to time-of-flight mass spectrometry (HPLC-TOF-MS). Isoliquiritigenin (ISO) had the strongest association with NAFLD out of the determinant metabolites. Individuals with higher concentrations of ISO had healthier metabolic and hepatic status and were less likely to have NAFLD (OR 0.13). Receiver operating characteristic (ROC) curves demonstrated the predictive power of ISO in panel combination with other NAFLD and IR-related variables, such as visceral adipose tissue (VAT) (AUROC 0.972), adiponectin (AUROC 0.917), plasmatic glucose (AUROC 0.817), and CK18-M30 (AUROC 0.810). Individuals with lower levels of ISO have from 71 to 82% more risk of presenting NAFLD compared to individuals with higher levels. Metabolites such as ISO, in combination with visceral adipose tissue, IR, and related markers, constitute a potential non-invasive tool to predict and diagnose NAFLD.