Purpose <p>Unstable Surface Training (UST) enhances postural control, yet its efficacy for transferable balance and optimal dosing remains debated. To quantify UST effects on balance in healthy individuals and establish evidence-based dose-response guidelines.</p> Methods <p>A systematic search of four databases identified 47 randomized controlled trials. A three-level random-effects meta-analysis was employed to account for data dependency. Sensitivity analyses were performed to identify influential cases and robust dose-response relationships.</p> Results <p>UST significantly improved overall static (Hedges’<i>g</i> = 0.55, <i>P</i> &lt; 0.01) and dynamic balance (<i>g</i> = 0.33, <i>P</i> &lt; 0.01). Subgroup analysis revealed extreme task specificity: gains were largest for functional tasks on unstable surfaces (<i>g</i> = 1.26) but significantly lower for stable-ground laboratory measures (<i>g</i> = 0.45). For static balance, an exploratory meta-regression suggested a non-linear trend peaking at 5–6 weeks; however, this observation should be interpreted cautiously due to study clustering. For dynamic balance, a frequency of 3 sessions/week yielded the highest certainty of improvement (<i>g</i> = 0.45).</p> Conclusion <p>UST is an effective modality for enhancing postural control under labile conditions. While limited transfer to stable surfaces suggests it should serve as a supplementary neural primer, the observed 5–6 week adaptation peak is strictly hypothesis-generating. Future longitudinal trials are required to validate this temporal pattern, whereas an optimal frequency of 3 sessions/week is currently supported for dynamic balance.</p> Graphical Abstract <p></p>