Abstract Objective The objective of this study was to quantify the effectiveness of virtual reality balance games (VRBG) to decrease risk and fear of falls among women. Methods Thirty six community dwelling women aged 56 and above were randomly divided into experimental (exercises using VRBG focus on improving balance) and control (conventional balance exercises) groups. Both groups attended a twice 6 weekly exercise session for an hour. Risk and fear of falls were measured with Physiological Profile Approach (PPA) and Activity Specific Balance Scale (ABC-6). Pre and post intervention differences between the groups were examined using two way repeated measures ANOVA. Results Both VRBG and conventional balance exercise groups had significant decrease in PPA ( p < 0.001) and ABC-6 ( p < 0.01) after the interventions. However, no significant effects were demonstrated between the groups in PPA ( p = 0.18) and ABC-6 ( p = 0.25) post intervention. Time and group interaction effect were not significant for PPA ( p = 0.18) and ABC-6 ( p = 0.45). Conclusions Practising VRBG can increase balance confidence and decrease risk of falls among community dwelling women.