Background and aims: Up to 40% of all dementia cases may be preventable, primarily by treating or acting on well-established cardiovascular risk factors such as diabetes, hypertension, smoking, and physical inactivity. Whether physical inactivity is associated with risk of non-Alzheimer's dementia – a disease influenced by cardiovascular risk factors – and whether a given association differs for physical activity in leisure time and at work remains unknown. Methods: We conducted a prospective cohort study including 117,616 individuals from the Copenhagen General Population Study and the Copenhagen City Heart Study with up to 43 years of follow-up. Results: Multifactorially adjusted hazard ratios for low versus high physical activity at leisure time was 1.60 (95% confidence interval 1.40–1.83) for non-Alzheimer's dementia and 0.94 (0.80–1.11) for Alzheimer's disease. Corresponding values for non-Alzheimer's dementia after additional adjustment for physical activity at work or apolipoprotein E (APOE) genotype were 1.60 (1.40–1.83) and 1.82 (1.34–2.15). Multifactorially and APOE adjusted hazard ratios for high versus low physical activity at work were 1.50 (1.10–2.05) for non-Alzheimer's dementia and 1.62 (1.14–2.31) for Alzheimer's disease. When combining the two types of physical activity, physical activity in leisure time had the strongest relationship with risk of non-Alzheimer's dementia. Conclusions: Physical inactivity in leisure time was associated with increased risk of non-Alzheimer's dementia, independent of modifiable risk factors and physical activity at work. The present study thus provides evidence for public health advice on physical activity in leisure time for the vascular part of dementia.