Falls occur mainly while walking or performing concurrent tasks. We determined whether a music-based multitask exercise program improves gait and balance and reduces fall risk in elderly individuals. We conducted a 12-month randomized controlled trial involving 134 community-dwelling individuals older than 65 years, who are at increased risk of falling. They were randomly assigned to an intervention group (n = 66) or a delayed intervention control group scheduled to start the program 6 months later (n = 68). The intervention was a 6-month multitask exercise program performed to the rhythm of piano music. Change in gait variability under dual-task condition from baseline to 6 months was the primary end point. Secondary outcomes included changes in balance, functional performances, and fall risk. At 6 months, there was a reduction in stride length variability (adjusted mean difference, -1.4%; P < .002) under dual-task condition in the intervention group, compared with the delayed intervention control group. Balance and functional tests improved compared with the control group. There were fewer falls in the intervention group (incidence rate ratio, 0.46; 95% confidence interval, 0.27-0.79) and a lower risk of falling (relative risk, 0.61; 95% confidence interval, 0.39-0.96). Similar changes occurred in the delayed intervention control group during the second 6-month period with intervention. The benefit of the intervention on gait variability persisted 6 months later. In community-dwelling older people at increased risk of falling, a 6-month music-based multitask exercise program improved gait under dual-task condition, improved balance, and reduced both the rate of falls and the risk of falling. Trial Registration clinicaltrials.gov Identifier: NCT01107288.