Aim To optimize vaccination strategy, evidence on vaccine efficacy against COVID-19 is needed. Method The present network meta-analysis uses reconstructed individual patient data from phase III trials on vaccine efficacy (VE), identified through MEDLINE, EMBASE, and Cochrane library (CENTRAL) peer-reviewed and published in English before August 31, 2021. The primary outcome was the VE against confirmed COVID-19 at any time after the first dose as defined in each study. VE was re-estimated using the two-stage approach. Poisson regression models were applied to each trial at the first stage, and the incidence risk ratio (IRR) and their 95% CI were aggregated to allow random-effects network meta-analysis (NMA) at the second stage. VE was expressed as: (1-IRR) × 100. The study protocol is registered in PROSPERO (CRD42020200012). Results A total of eight studies, evaluating nine different vaccines were identified and analyzed. Between April 23, 2020 and January 05, 2021, 210,418 participants were recruited in 354 sites worldwide. During a median (IQR) follow-up duration of 69.8 (69.7–70.3) days, 2131 confirmed COVID-19 cases occurred (604; 26.0 per 1000 person–years in vaccine recipients and 1527; 85.9 per 1000 person–years in the control group). The mRNA-1273 vaccine was the most effective (P-score 0.99); at any time after dose 1, incidence reduction for mRNA-1273 ranged from 78% to 98% compared to the other vaccines. Conclusion Our results provide evidence for the short-term superiority of mRNA vaccines, especially the mRNA-1273 vaccine in prevention of COVID-19 in different populations. Supplementary Information The online version contains supplementary material available at 10.1007/s10389-022-01707-1.