Background HbA1c, the most commonly used indicator of chronic glucose metabolism, is closely associated with cardiovascular disease. However, the relationship between HbA1c and the mortality of acute coronary syndrome (ACS) patients has not been elucidated yet. Here, we aim to conduct a systematic review assessing the effect of HbA1c on in-hospital and short-term mortality in ACS patients. Methods Relevant studies reported before July 2019 were retrieved from databases including PubMed, Embase, and Central. Pooled relative risks (RRs) and the corresponding 95% confidence interval (CI) were calculated to evaluate the predictive value of HbA1c for the in-hospital mortality and short-term mortality. Results Data from 25 studies involving 304,253 ACS patients was included in systematic review. The pooled RR of in-hospital mortality was 1.246 (95% CI 1.113–1.396, p: 0.000, I2 = 48.6%, n = 14) after sensitivity analysis in studies reporting HbA1c as categorial valuable. The pooled RR was 1.042 (95% CI 0.904–1.202, p: 0.57, I2 = 82.7%, n = 4) in random-effects model for studies reporting it as continuous valuable. Subgroup analysis by diabetic status showed that elevated HbA1c is associated increased short-term mortality in ACS patients without diabetes mellitus (DM) history and without DM (RR: 2.31, 95% CI (1.81–2.94), p = 0.000, I2 = 0.0%, n = 5; RR: 2.56, 95% CI 1.38–4.74, p = 0.003, I2 = 0.0%, n = 2, respectively), which was not the case for patients with DM and patients from studies incorporating DM and non-DM individuals (RR: 1.16, 95% CI 0.79–1.69, p = 0.451, I2 = 31.9%, n = 3; RR: 1.10, 95% CI 0.51–2.38), p = 0.809, I2 = 47.4%, n = 4, respectively). Conclusions Higher HbA1c is a potential indicator for in-hospital death in ACS patients as well as a predictor for short-term mortality in ACS patients without known DM and without DM.