OBJECTIVE: The objective of this study was to evaluate the experimental effects of exercise on memory function among adults with mild cognitive impairment (MCI). METHODS: A systematic review and meta-analysis were employed. Studies were identified using electronic databases, including PubMed, PsychInfo, SPORTDiscus, and Google Scholar. To be eligible for inclusion, studies had to employ an experimental design, be conducted in humans (no animal models) with mild cognitive impairment, include an exercise intervention arm (either acute exercise or chronic training), and measure memory function (any type) as the outcome measure. Comprehensive Meta-Analysis software was used to compute the standardized mean difference effect size (Cohen's d) and 95% CI using a random-effects model. RESULTS: In total, 11 studies met our criteria, contributing to 20 effect size estimates. Among the 20 effect size estimates, 13 contributed to a short-term memory recall, with 7 contributing to a delayed/long-term memory recall. The overall weighted mean effect size was d = 0.30 (95% CI: 0.16-0.44; P < 0.001), indicating a small to medium effect of exercise on improving memory function. For the moderation analysis, there was no evidence of a moderation effect for recall duration (Q = 0.004, df(1), P = 0.95), as exercise (vs. control) demonstrated both short-term (d = 0.31; 95% CI: 0.14-0.47; P < 0.001) and long-term memory (d = 0.29; 95% CI: 0.06-0.53; P = 0.013) benefits. CONCLUSION: Current evidence suggests that exercise may help to improve memory function among those with MCI. Safe and progressive forms of exercise should be promoted among MCI patients.