Virtual reality therapy (VRT) has clinical indications in rehabilitation programs for the elderly; however, there is still no consensus on the recovery of body balance. The objective of this review was to summarize the effects of physical therapy interventions with VRT in the rehabilitation of balance in the elderly. The studies were identified via a systematic search in the databases PubMed, SciELO, LILACS and PEDro from 2010 onward. Clinical trials with interventions that involved VRT in the elderly were included in the study and were subjected to methodological quality analysis using the PEDro scale. A random effects meta-analysis of the studies that analyzed balance using the Berg Balance Scale and the Timed Up and Go (TUG) test was performed. Ten articles met the inclusion criteria, which presented variability in relation to the types of interventions used (70%) and the outcomes analyzed (60%). The mean duration of the interventions was 13.90 (+/- 5.08) weeks, with at least two weekly sessions (+/- 0.73). There were positive results in relation to improvements in both dynamic and static balance (70% of the studies), mobility (80%), flexibility (30%), gait (20%) and fall prevention (20%). A summary of the meta-analysis showed mean effects on the Berg scale (standardized mean difference [SMD]: -0.848; 95% CI: -1.161; - 0.535) and the TUG test (SMD: 0.894; 95% CI: 0.341; 1.447). Individually, virtual reality is promising in rehabilitation programs for the elderly. The overall measures were sufficient to show beneficial effects of the therapy on balance in the elderly.