Objective: To identify the dietary patterns of individuals with metabolic syndrome and determine associations with sociodemographic, lifestyle, anthropometric and biochemical features. Methods: We analyzed data of individuals with metabolic syndrome attending the nutrition outpatient clinic at a public university in São Paulo, from 2008 to 2011. We assessed food intake by 24-hour dietary recall and identified the dietary patterns using principal components analysis. Multivariate multiple linear regression models were applied to investigate possible associations between dietary patterns and the independent variables. We applied Pearson's correlations to evaluate possible associations between dietary patterns and biochemical and anthropometric variables. Results: Of the 166 outpatients enrolled, 65.7% (109) were women; 63.0% (103) were adults (21-59 years) and 37.0% were elderly (60-79 years). Most people were married, sedentary and obese, with incomplete primary education, non-smoking and non-alcoholic. We identified nine dietary patterns. There was a positive association between the "traditional Brazilian" dietary pattern (rice, beans and red meat) with men and a negative association with age, and the "prudent" (whole grains, soups) with women. There was a positive correlation between the "traditional Brazilian", "compensatory" (snacks, sweeteners, light products) and "mixed" (refined grains, fruits, oils) patterns with fasting glucose, and negative correlation between the "healthy" (vegetables and oilseeds) and waist circumference. Conclusion: The results of this study indicate that these individuals with metabolic syndrome have dietary patterns characterized by the presence of refined foods, fats, red meat and alcohol, associated with biochemical disturbance, as impaired fasting glucose and hypertriglyceridemia.