To determine whether physical exercise enhances cognition following TBI or stroke. Studies were identified through searches of PubMed, ScienceDirect and the reference lists of papers that were included for full-text evaluation. Medical subject headings from three concepts, i.e. brain injury, physical exercise and cognition, were used to incorporate related search terms. Included were all trials published in English that assessed cognition before and after an exercise intervention in human adults with TBI or stroke. Nine randomized and two non-randomized controlled trials, as well as three single group pre-post studies were included. Relevant data concerning the methods and results of the included studies were extracted. Methodological quality of the RCT's was evaluated using the PEDro scale. Non-randomized trials were assessed using the Downs and Black checklist. The included trials were generally of medium methodological quality, though often plagued with issues of internal and external validity. The studies exhibited great heterogeneity, rendering a meta-analysis infeasible. Though well-designed studies are still needed, the preponderance of evidence suggests a positive effect of physical exercise on global cognitive functioning, especially in the chronic stages of a brain injury. Time after injury as well as the duration of the exercise program are mediating factors.