INTRODUCTION & AIMS Exercise training trials consistently show efficacy for improving cognitive outcome in older adults, especially when the intervention is supervised. However, whether this improvement in cognition is sustained long-term (i.e. a legacy effect) is unknown. While this is likely impacted by long-term adherence to exercise, the legacy effect of exercise training trials on cognition has not been established. Therefore, the aim of this study was to review the literature and investigate the legacy effect of an exercise trial on cognitive function in older adults. METHODS A systematic review of the literature published in English was conducted in PubMed, Scopus, CINAHL, SPORTdiscus, Web of Science and Cochrane Central Register of Controlled Trials electronic databases. Study screening and data extraction were undertaken by two independent researchers. The main inclusion criteria were randomised controlled trial design with non-exercise control, older adult cohort (average age ≥50 years), supervised exercise intervention and cognitive assessments at post-trial and follow up (≥3 months). Pairwise random-effects restricted maximum likelihood meta-analysis estimated the standardised mean difference (Hedges’ g) between intervention and control groups for each cognitive function outcomes at post-trial and follow-up. RESULTS A total of 12,539 articles were screened with 21 trials (n=2,504 participants aged 75.2±6.4 and mostly (56%) female) included in a meta-analysis. After 5.2±3.2 months, a legacy effect was seen with a small-to-moderate between-group effect for the cognitive domain of executive function favouring the exercise interventions (SMD=0.30, 95% CI 0.11 to 0.49; p=0.002). No effect was observed for other cognitive domains. The Cochrane risk-of-bias tool for randomized trials (RoB2) tool identified that 73.7% of the trials had “some concerns” and 26.3% had high risk for bias. CONCLUSION In older adults who received an exercise intervention there was a small-to-moderate legacy effect for improved executive function compared to controls, five months after trial completion.