In this work, we report a systematic review and meta-analysis that seeks to analyze the accuracy of diagnostic tests for coronavirus disease 2019 (COVID-19). The objective of this article is to detail the scientific findings based on diagnostic tests of the last years when the pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) occurred. Searches for published studies were carried out in the PubMed database between the years 2020 and 2021 for the diagnosis of COVID-19. Ninety-nine scientific articles that met the criteria were examined and accepted in the meta-analysis, and the diagnostic accuracy was evaluated through specificity and sensitivity. Molecular tests [Reverse transcription polymerase chain reaction (RT-PCR), reverse transcription loop-mediated isothermal amplification (RT-LAMP), and clustered regularly interspaced short palindromic repeats (CRISPR)] showed better performance in terms of sensitivity and specificity when compared to serological tests [Enzyme-linked immunosorbent assay (ELISA), chemiluminescence immunoassay (CLIA), lateral flow immunoassay (LFIA), chemiluminescent microparticle immunoassays (CMIA), and Fluorescence immunoassay (FIA)], which showed higher specificity, mainly for the detection of IgG antibodies; however, they showed sensitivity <90%. In addition, the antiviral neutralization bioassay (ANB) diagnostic test demonstrated high potential for the diagnosis of COVID-19, since it obtained the highest area under the curve restricted to the false-positive rates (AUCFPR) of 0.984. It is settled that the different diagnostic tests have been efficiently adapted for the detection of SARS-CoV-2; however, their performance still needs to be optimized to control future outbreaks of COVID-19, which will also serve to help the control of future infectious agents.