Non-alcoholic fatty liver disease is a rapidly growing public health problem. In this study, we explored the association between dietary patterns (DPs) and fatty liver and liver function tests. This was a cross-sectional study using data from the US community-based National Health and Nutrition Examination Survey. Participants with data on dietary intake, blood pressure, and status for diabetes mellitus were analyzed. DPs were determined by principal components analysis. Analysis of covariance and logistic regression models accounted for the survey design and sample weights. Of the 20643 eligible participants, 45.7% had prevalent fatty liver. Three DPs collectively explained 50.8% of variance in dietary nutrients consumption. The first DP was representative of a diet containing high levels of saturated and mono-unsaturated fatty acids, total fat and carbohydrate; the second DP comprised vitamins, minerals and dietary fibre; and the third DP was mainly representative of polyunsaturated fatty acids. In adjusted multivariable regression models, participants in the top quarter of the second DP had 34% lower odds of prevalent fatty liver (odds ratio 0.66 [95% confidence interval [CI]: 0.43-0.71]), while those in the top quarter of the first DP had 86% higher odds (1.86 [95% CI: 1.42-2.95]) of prevalent fatty liver, relative to participants in the bottom quarter of each of the DPs. Our findings suggest that a diet with high load of vitamins, minerals, and fiber content is associated with a lower prevalence of non-alcoholic fatty liver disease.