Falls among people living with dementia are a major adverse health outcome, strongly associated with physical disability, decline in activities of daily living (ADL), institutionalization, and increased mortality risk. The incidence of falls in this population is consistently higher than in older adults with preserved cognitive function. Although exercise interventions centered on lower-limb strength and balance training have been firmly established as effective for reducing falls among community-dwelling older adults, standard fall-prevention programs such as the Otago Exercise Programme (OEP) are implicitly designed under the assumption that participants retain adequate comprehension, memory, executive function, and self-management capacity. As a result, these programs are prone to implementation failure in dementia care and clinical settings.This paper aims to theoretically reconfigure fall-prevention exercise by decomposing existing evidence into “core active ingredients” and “design elements that impose excessive burden in dementia,” and by reconstructing a dementia-adapted framework for fall-prevention exercise. Specifically, we propose a Dementia-adapted Otago Exercise Programme (D-OEP) based on four core principles: (1) radical simplification of task structure; (2) exclusion of high-risk static balance tasks; (3) embedding balance stimuli within functional movements such as sit-to-stand and supported ankle exercises; and (4) delivery formats that assume caregiver supervision and support.Rather than eliminating balance training, this framework repositions balance stimuli into safe, repeatable functional activities, thereby suppressing fear responses and maladaptive reactions while ensuring cumulative exposure to lower-limb strength and postural control demands. The value of fall-prevention exercise lies not in theoretically optimal prescriptions but in cumulative exposure to active ingredients achieved through feasibility, adherence, and safety. This paper reframes fall prevention not as an issue of “exercise inefficacy” but as a problem of design and implementation, and provides a conceptual foundation for translating evidence into dementia care practice.