Background: Overexpression of Aurora B is linked to poor prognosis in various malignancies; however, its prognostic role remains debatable. Conducting a meta-analysis is essential to reach a definitive conclusion. Methods: Various databases were searched. Aurora B protein expression was assessed for prognostic significance using pooled hazard ratio (HR) with a 95% confidence interval (CI). Meta-regression and subgroup analysis identified the source of heterogeneity. Results: The study comprised 1384 cancer patients from 10 articles. The result with multivariate data (pooled HR=1.18, 95% CI= 0.71-1.95, p=0.52, I²=83%) and univariate data (pooled HR=1.81, 95% CI=0.92, 3.57, p=0.09, I2=89%) showed that increased Aurora B expression was not linked with poor overall survival (OS). Subgroup analysis showed that sample size, follow-up time, cut-off value, non-Chinese patients, antibody source were not associated with unfavorable OS. Conclusions: Aurora B expression could not be used as a prognostic marker in cancer.