A prime–boost strategy of COVID-19 vaccines brings hope to limit the spread of SARS-CoV-2, while the immunogenicity of the vaccines is waning over time. Whether a booster dose of vaccine is needed has become a widely controversial issue. However, no published meta-analysis has focused on the issue. Therefore, this study assessed the immunogenicity and safety of the different combinations of prime–boost vaccinations. Electronic databases including PubMed, the Cochrane Library, Embase, medRxiv, Wanfang and CNKI were used to retrieve the original studies. A total of 28 studies, 9 combinations of prime–boost vaccinations and 5870 subjects were included in the meta-analysis, and random effect models were used to estimate pooled immunogenicity and safety. The immunity against COVID-19 after the prime vaccination waned over time, especially in the populations primed with inactivated vaccines, in which the seropositive rate of antibodies was only 28% (95% CI: 17–40%). Booster vaccination could significantly increase the antibody responses, and heterologous immunization was more effective than homologous immunization (neutralization titers: 1.65 vs. 1.27; anti-RBD IgG: 1.85 vs. 1.15); in particular, the combination of inactivated–mRNA vaccines had the highest antibody responses (neutralization titers: MRAW = 3.64, 95% CI: 3.54–3.74; anti-RBD IgG: 3.73, 95% CI: 3.59–3.87). Moreover, compared with the initial two doses of vaccines, a booster dose did not induce additional or severe adverse events. The administration of the booster dose effectively recalled specific immune responses to SARS-CoV-2 and increased antibody levels, especially in heterologous immunization. Considering the long-term immunogenicity and vaccine equity, we suggest that now, only individuals primed with inactivated vaccines require a booster dose.