Background: This meta-analysis aimed to determine whether exercise interventions were effective in preventing fall-related fractures in older people. The treatment effects on rate of falls, leg strength and balance were also examined. Methods An electronic database search was conducted in PubMed, EMBASE, the Cochrane library and PEDro up to 1 September 2015. Randomized controlled trials (RCTs) that conducted exercise interventions and reported fall-related fracture data in older people were included. The primary outcome was the treatment effects on fall-related fractures determined by relative risk (RR) and 95% confidence interval (CI). The treatment effects on falls, leg strength and balance were also reported using rate ratio (RaR) with 95% CI and standardized mean difference (SMD) with 95% CI, respectively. Random effects models were used for meta-analysis. Results: Fifteen studies including 3136 participants met the inclusion criteria. Exercise had a beneficial effect on reduction of fall-related fractures, with pooled estimates of RR 0.604 (95% CI 0.453 - 0.840, P ¼ 0.003, I2¼ 0%). The rate of falls (RaR 0.856, 95% CI 0.778 – 0.941, P ¼ 0.001, I2¼ 45%) and leg strength (SMD 0.613, 95% CI 0.119 - 1.107, P ¼ 0.015, I 2¼ 76.7%) were also potentially affected by exercise interventions. These only had a marginally beneficial effect on balance (SMD 0.468, 95% CI 0.011 – 0.947, P ¼ 0.055, I 2¼ 93.6%). Conclusions: Our findings implied that exercise interventions were effective in preventing fall-related fractures and reducing risk factors of fall-related fractures in older people.