Physical activity has the potential to improve health outcomes in people with dementia, namely when living at home. However, the knowledge about home-based physical activity for this population is scarce. Thus, we aim to identify and synthesize the effects of home-based physical activity for people with dementia. A systematic review was conducted. Quality of studies was assessed using the Delphi List. Effect sizes (ES) were calculated with MetaXL 2.0. A meta-analysis was conducted for the Mini-Mental Status Examination (MMSE), Neuropsychiatric Inventory (NPI), Cornell Scale for Depression in Dementia, Alzheimer's Disease Cooperative Study Group Activities of Daily Living Scale (ADCS-ADL), Functional Reach test, Timed Up and Go (TUG) test, Short Physical Performance Battery, Dementia Quality of Life, NPI Caregivers subscale and Zarit Burden Interview (ZBI). Sixteen randomized controlled trials were included. Most were of high quality and published after 2015. A large heterogeneity of interventions was found. Meta-analysis showed significant results in MMSE (ES = 0.71, 95% CI 0.43, 0.99), NPI (ES = -0.37, 95% CI -0.57, -0.17), ADCS-ADL (ES = 0.80, 95% CI 0.53, 1.07), Functional Reach test (ES = 2.24, 95% CI 1.80, 2.68), TUG test (ES = -2.40, 95% CI -2.84, -1.96), NPI Caregivers subscale (ES = -0.63, 95% CI -0.94, -0.32), and ZBI (ES = -0.45, 95% CI -0.77, -0.13). Few minor adverse events and high adherence to intervention were found. Home-based physical activity seems safe and effective in delaying cognitive function decline and improving changes in behavioral and psychological symptoms of dementia, activities of daily living, health-related physical fitness, and carer's burden in people with dementia living at home.