Background: Reactive balance is the last line of defense to prevent a fall when the body loses stability, and beneficial effects of various exercise-based interventions on reactive balance in older adults have been reported. However, their pooled evidence on the relative effects has yet to be described. Objective: To review and evaluate the comparative effectiveness of various exercise-based interventions on reactive balance in older adults. Methods: Nine electronic databases and reference lists were searched from inception to August 2021. Eligibility criteria according to PICOS criteria were as follows: (1) population: older adults with the mean age of 65 years or above; (2) intervention and comparison: at least two distinct exercise interventions or one exercise intervention with no-exercise controlled intervention (NE) compared in each trial; (3) outcome: at least one measure of reactive balance; (4) study: randomized controlled trial. The main network meta-analysis was performed on data from the entire older adult population, involving all clinical conditions as well as healthy older adults. Subgroup analyses stratified by characteristics of participants (healthy only) and reactive balance outcomes (simulated slip or trip while walking, simulated forward falls, being pushed or pulled, and moveable platform) were also conducted. Results: Thirty-nine RCTs (n = 1388) investigating 17 different types of exercise interventions were included in the network meta-analysis. Reactive balance training as a single intervention presented the highest probability (surface under the cumulative ranking (SUCRA) score) of being the best intervention for improving reactive balance and the greatest relative effects vs. NE in the entire sample involving all clinical conditions [SUCRA = 0.9; mean difference (95% Credible Interval): 2.7 (1.0 to 4.3)]. The results were not affected by characteristics of participants (i.e., healthy older adults only) or reactive balance outcomes. Summary/Conclusion: The findings from the NMA suggest that a task-specific reactive balance exercise could be the optimal intervention for improving reactive balance in older adults, and power training can be considered as a secondary training exercise.