The purpose of this study was to investigate the relationship between slow usual walking speed and all-cause mortality risk in older people by conducting a meta-analysis. We searched through the Pubmed, Embase and Cochrane Library database up to March 2015. Only prospective observational studies that investigating the usual walking speed and all-cause mortality risk in older adulthood approaching age 65 years or more were included. Walking speed should be specifically assessed as a single-item tool over a short distance. Pooled adjusted risk ratio (RR) and 95% confidence interval (CI) were computed for the lowest versus the highest usual walking speed category. A total of 9 studies involving 12,901 participants were included. Meta-analysis with random effect model showed that the pooled adjusted RR of all-cause mortality was 1.89 (95% CI 1.46–2.46) comparing the lowest to the highest usual walk speed. Subgroup analyses indicated that risk of all-cause mortality for slow usual walking speed appeared to be not significant among women (RR 1.45; 95% CI 0.95–2.20). Slow usual walking speed is an independent predictor of all-cause mortality in men but not in women among older adulthood approaching age 65 years or more.