The paper discussed the main factors that explain the Brazilian agricultural exports. In order to achieve this goal, the paper applied a gravity model that includes fixed and random effects estimations, besides the Ordinary Least Squares (OLS) approach. Distance, trade partners´ GDP, and geographical localization were the significant variables. Moreover, puzzle effects are associated to exchange rate, partners´ agricultural exports profile and the partners´ agricultural share in GDP. Finally, this study highlights the potential change of the relevant variables because of specific characteristics of each commercial Brazilian partner.