Background and Aims: This study aims to compare the outcomes of early tracheostomy (ET) (<= 10 days after translaryngeal intubation) with late tracheostomy (LT) (>10 days after translaryngeal intubation) in critically ill patients with prolonged mechanical ventilation (MV). Methods: We searched PubMed, EMBASE and the Cochrane Library from inception to April 2014. We included all randomized controlled trials (RCTs), which compared ET with LT in critically ill patients. There was no language restriction. Two authors extracted data and conducted a quality assessment. Meta-analyses using the fixed-effects or random-effects model were conducted for mortality, incidence of ventilator-associated pneumonia (VAP), duration of MV and sedation, length of intensive care unit (ICU) stay. Results: We enrolled 9 studies, in which a total of 2040 patients were randomized to either ET group (N = 1018) or LT group (N = 1022). ET might reduce the duration of sedation [weighted mean difference (WMD) = -5.99 days; 95% confidence intervals (CI) = -11.41 to -0.57 days; P = 0.03]. ET did not significantly alter the mortality [relative risk (RR) = 0.88; 95% CI = 0.76-1.00; P = 0.06], incidence of VAP (RR = 0.84; 95% CI = 0.66-1.08; P = 0.17), duration of MV (WMD = -4.46 days; 95% CI = -12.61 to 3.69 days; P = 0.28) and length of ICU stay (WMD = -7.57 days; 95% CI = -15.42 to 0.29 days; P = 0.06). Conclusions: Our meta-analysis suggested that ET might be able to reduce the duration of sedation but did not significantly alter the mortality, incidence of VAP, duration of MV and length of ICU stay.