Objective: Brain-derived neurotrophic factor (BDNF) has been hypothesized as a potential mechanism through which exercise may subserve memory function. The present review specifically evaluates this hypothesis. Methods: Studies were identified using electronic databases, including PubMed, PsychInfo, Sports Discus and Google Scholar. Results: In total, 52 articles met the study criteria, and among these, 36 were conducted in an animal model and 16 among humans. Among the animal experiments, 100% of them demonstrated that chronic exercise improved memory function; 97% demonstrated an exercise-induced increase in BDNF; and among the eight evaluating BDNF as a mediator, 100% provided evidence that BDNF mediated the relationship between exercise and memory. The findings in the human studies were mixed. Among the human studies, 44% demonstrated that varying exercise protocols improved memory and increased BDNF levels, and among the studies evaluating BDNF as a mediator, 40% provided evidence that BDNF mediated the relationship between exercise and memory. Conclusion: In animal models, chronic exercise training robustly increases BDNF and improves memory performance, with reasonable evidence to also suggest that BDNF may mediate the exercise-memory interaction. These interrelationships, however, are less clear among humans. Future research among humans, in particular, is needed to evaluate the extent to which BDNF may mediate the relationship between exercise and memory.