Ethnobiological investigations have focused on identifying factors that interfere with the criteria adopted for selection of plants, especially medicinal plants, by different populations, confirming the theory that plant selection is not random. However, regarding wild food plants, little effort has been made to confirm the theory in this context, especially in Brazil. Therefore, this systematic review aimed to contribute to the establishment of theoretical bases of the non-random selection of wild food plants by local populations in Brazil. For this, searches were made in 4 databases, namely, Web of Science, Scielo, Scopus and PubMed, using 8 sets of keywords in English and Portuguese in order to identify wild food plants occurring in Brazil. The steps were: application of inclusion and exclusion criteria, screening of articles, selection of studies based on risk of bias, data treatment and, finally, data analysis. Eighty articles met the inclusion criteria of this review. However, 45 of them were considered to present high risk of bias and thus 35 articles were kept for the identification of overused and underused families. The results were inferred through two different approaches (IDM and Bayesian). Annonaceae, Arecaceae, Basellaceae, Cactaceae, Capparaceae, Caryocaraceae, Myrtaceae, Passifloraceae, Rhamnaceae, Rosaceae, Sapotaceae, Talinaceae, and Typhaceae were considered overused. Eriocaulaceae, Orchidaceae, and Poaceae were considered underused. Therefore, considering that some families are more (or less) used than others, we confirm that the wild food plants occurring in Brazil, known and used by different populations, are not chosen at random.