Aim: Dual-task methods, in which walking is the primary task, are not sufficient for accurately screening for the risk of falls among healthy older adults. Therefore, the goal of this research was to investigate whether using a dual-task method over an extended walking distance can predict falls among community-dwelling older adults. Methods: We enrolled independent community-dwelling adults aged ≥65 years. Physical performance, cognitive function, psychological function, and a dual-task test were assessed at baseline. Our dual-task test required the subjects to walk 60 m while stepping over lines. The intervals between the lines ranged from 50–100 cm and were unequal. Falls and fall-related injuries were measured over a 12-month follow-up period using monthly postal surveys. Results: Ninety-two of 118 subjects (mean age, 75.4±5.5 years) completed the 12-month follow-up. Sixteen (17.4%) of fallers had injurious falls or fell more than or equal to two times. There were no significant differences between the fallers and non-fallers, except in age and in the number of missteps during the dual-task test when walking ≥40 m. The Kaplan–Meier analysis revealed that those who had more than one misstep while walking ≥40 m had a significantly higher incidence of injurious or multiple falls than those who had no missteps. Conclusion: Our findings suggest that the dual-task method with an extended walking distance may be able to predict falls among community-dwelling older adults.