Background: Despite increased interest in combining learning and physical activity (PA), the academic and PA benefits of active learning are uncertain. Methods: A systematic search of 5 databases for studies combining learning math with PA in primary/elementary schools was conducted. Academic benefit was evaluated by pre-post intervention math scores compared to a control group. Effect sizes (ES) were extracted/calculated when possible. Due to study heterogeneity, meta-analysis was not conducted. Results: Six randomized controlled trials and 5 quasi-experimental studies evaluating 4082 participants (53% girls; mean age 7.5-11.1 years) were eligible. Math scores were significantly better in the intervention group in 6 of 11 studies on at least 1 test (ES: 0.42-4.7; p[less than or equal to]0.03). Other math tests either were not all statistically significant (2 studies) or the benefit varied across grades (1 study). No studies reported a decline in math scores. Of studies measuring PA with accelerometers, 4 of 5 reported significantly greater PA in the intervention group during the intervention (p < 0.05) or across the school day (p < 0.01). Conclusions: Undertaking PA while learning was largely equivocal for math scores but showed promising results for increasing daily PA, without detrimental effects on math performance. The need for more rigorous studies with comprehensive assessment of academic performance and PA is highlighted.