A large body of epidemiological and experimental data exploring the relationship between physical activity (PA) and Alzheimer's disease (AD) are now available. Despite observational evidence supporting a role for PA in delaying the onset of AD, randomised controlled trials have reported mixed findings, likely due to the heterogeneity in study cohorts, outcome measures, and the adopted PA intervention. The primary objective of this narrative review is to evaluate the extant evidence on the relationship between PA, cognitive decline and AD in older populations. The interaction between PA and the putative mechanisms underlying AD progression, including genetic factors and amyloid-β levels will be explored. In this context, particular attention will be given to studies assessing PA in the early clinical and preclinical, asymptomatic stages of AD. Based on current evidence, clinical considerations for implementation of exercise-based interventions are discussed, along with limitations of previous research and directions for future studies.