QUESTION: Can exercise or physical training improve physical function and prevent falls in older adults with visual impairments? DESIGN: Systematic review of randomised controlled trials with meta-analysis. PARTICIPANTS: Older adults (≥ 60 years) with visual impairments. INTERVENTION: Individual or group exercise or physical training classes in any settings. OUTCOME MEASURES: Mobility, balance, strength and proprioception measured with performance tests or questionnaires and/or falls with calendars or incident reports. RESULTS: Four eligible trials with a total of 522 participants were identified. Multimodal group exercise (n = 50 and 41) and Tai Chi (n=40) improved physical function among residents of care settings. Meta-analysis of data from two trials indicated a significant positive impact of multimodal exercise on the Berg Balance Score (weighted mean difference 3.9 points, 95% CI 1.8 to 6.0), but not on the Timed Up and Go test (weighted mean difference 1.5seconds, 95% CI -1.7 to 4.6). One trial (n=41) found that multimodal exercise reduced the time to first fall (p=0.049). A factorial trial (n=391) among community dwellers did not find a significant effect on falls from a home-based exercise intervention, although clinically relevant effects in either direction were not excluded by the study (incidence rate ratio=1.15, 95% CI 0.82 to 1.61). CONCLUSION: Exercise interventions in residential care settings improve performance on some tests of physical function that are risk factors for falls but the impact on falls is not yet clear. The impact of exercise and training on physical function and falls in community-dwelling older adults with visual impairments also warrants further investigation.