Aim The purpose of the present study was to investigate whether sarcopenia was associated with future falls in the general Japanese older population. Methods This study was a 2‐year prospective observational study. Participants were recruited from individuals who had an annual town‐sponsored medical check‐up and had not received nursing care. The inclusion criteria for participants in our study were: (i) agreement to participate; (ii) living independently; and (iii) the ability to walk to where the survey was carried out and to provide self‐reported data. A total of 223 residents (82 men, 141 women) participated in the baseline assessment in the study. Demographic information, previous fall history, locomotive syndrome, body function and structural measurements and pain at the knee and/or lumber spine were assessed. The Asian Working Group for Sarcopenia algorithm was used to classify the presence of sarcopenia, and assess the history of falling when the participant received their annual medical check‐up. Results A total of 162 participants had an annual follow‐up assessment, 50 of whom (30.8%) fell at least once during the 2‐year observational period after baseline assessment. Previous falling history, prevalence of locomotive syndrome, sarcopenia and pain were significantly higher in participants who had fallen compared with participants who had not. Multiple logistic regression analysis showed the prevalence of sarcopenia was a significant predictor of falling. Conclusion The key finding of the present study suggests that sarcopenia is a risk factor for falling in older adults who are living independently even after adjustment for previous falls and confounding factors. Geriatr Gerontol Int 2017; ••: ••–••.