Background: Health behavior theories are useful to interpret adolescents' dietary behaviors. Evidences show that theories are influenced by social and psychological determinants. So, the purpose of this study is to systematically review studies that tested social cognitive theories (non-integrated) that predict adolescents' dietary behaviors. Methods: A structured electronic search of all publication years (through April 2016) was conducted to identify studies in MEDLINE, SciELO, PsycINFO, Scopus, and LILACS with full text. Included publications were cross-sectional and longitudinal (non-intervention) studies involving adolescents (10 to 18 years) that examined the associations between constructs of social-cognitive theories and dietary behaviors. Related strings in titles, abstracts, and indexing fields were searched. Results: Theories used to explain dietary intake were the planned behavior and the social cognitive. It was observed evidences of positive associations between the social cognitive constructs and the fruits, the vegetables, the milk groups, and the whole-wheat foods (e.g., bread rich in fiber) and negative associations with sugarsweetened beverages, soft drinks, snacks high in fat, sugar, and/or sodium, and sweet treats. Theories explained greater proportion of variance for intention to dietary intake. The variance for intention ranged from 3% for pizzas, candy bars, candies, and sugar-sweetened beverages to 68% for whole-wheat food (i.e., bread rich in fiber). Conclusion: Longitudinal designs are necessary to comprehend the theories and evaluate the behavioral changes. Finally, the use of food groups should be employed in the studies to help the comparisons and present higher reproducibility. Studies always based on objective, systematic, and rigorous evidences.