BACKGROUND & AIMS: Our aims were to develop a method to accurately predict non-alcoholic fatty liver disease (NAFLD) and liver fat content based on routinely available clinical and laboratory data and to test whether knowledge of the recently discovered genetic variant in the PNPLA3 gene (rs738409) increases accuracy of the prediction. METHODS: Liver fat content was measured using proton magnetic resonance spectroscopy in 470 subjects, who were randomly divided into estimation (two thirds of the subjects, n = 313) and validation (one third of the subjects, n = 157) groups. Multivariate logistic and linear regression analyses were used to create an NAFLD liver fat score to diagnose NAFLD and liver fat equation to estimate liver fat percentage in each individual. RESULTS: The presence of the metabolic syndrome and type 2 diabetes, fasting serum (fS) insulin, FS-aspartate aminotransferase (AST), and the AST/alanine aminotransferase ratio were independent predictors of NAFLD. The score had an area under the receiver operating characteristic curve of 0.87 in the estimation and 0.86 in the validation group. The optimal cut-off point of -0.640 predicted increased liver fat content with sensitivity of 86% and specificity of 71%. Addition of the genetic information to the score improved the accuracy of the prediction by only <1%. Using the same variables, we developed a liver fat equation from which liver fat percentage of each individual could be estimated. CONCLUSIONS: The NAFLD liver fat score and liver fat equation provide simple and noninvasive tools to predict NAFLD and liver fat content.