BACKGROUND: Regular physical activity has multiple health benefits, especially in older people. Therefore, the World Health Organization recommends at least 2.5 h of moderate physical activity per week. The aim of the POWER Study was to investigate whether volunteer-assisted walking improves the physical performance and health of older people. METHODS: We approached people aged 65 years and older with restricted mobility due to physical limitations and asked them to participate in this multicentre randomised controlled trial. The recruitment took place in nursing homes and the community setting. Participants randomly assigned to the intervention group were accompanied by volunteer companions for a 30-50 min walk up to three times a week for 6 months. Participants in the control group received two lectures that included health-related topics. The primary endpoint was physical function as measured with the Short Physical Performance Battery (SPPB) at baseline and 6 and 12 months. The secondary and safety endpoints were quality of life (EQ-5D-5L), fear of falling (Falls Efficacy Scale), cognitive executive function (the Clock Drawing Test), falls, hospitalisations and death. RESULTS: The sample comprised 224 participants (79% female). We failed to show superiority of the intervention with regard to physical function (SPPB) or other health outcomes in the intention-to-treat analyses. However, additional exploratory analyses suggest benefits in those who undertook regular walks. The intervention appears to be safe regarding falls. CONCLUSIONS: Regular physical activity is essential to preserve function and to improve health and quality of life. Against the background of a smaller-than-planned sample size, resulting in low power, and the interference of the COVID-19 pandemic, we suggest that community based low-threshold interventions deserve further exploration. TRIAL REGISTRATION: The trial was registered with the German Clinical Trials Register ( www.germanctr.de ), with number DRKS00015188 on 31/08/2018.