Infectious disease coronavirus 2019 (COVID-19) is caused by the SARS-CoV-2 virus, and it mainly affects the upper respiratory tract. The gold standard for its diagnosis is real-time reverse transcription polymerase chain reaction (RT-qPCR) performed on a nasopharyngeal swab. In contrast, testing saliva has significant advantages as a diagnostic method. We searched for articles evaluating saliva as a diagnostic method for COVID-19 on the PUBMED/MEDLINE, WEB OF SCIENCE, COCHRANE, and SCIELO platforms. We initially found 233 articles and 20 were selected for inclusion following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses protocol: 18 cross-sectional studies and 2 case reports, including 8 from America, 8 from Asia, and 4 from Europe. The studies evaluated the presence of viral RNA, IgG, IgM, and IgA in samples of unstimulated saliva from adults with confirmed or suspected COVID-19. The vast majority of the studies performed RT-qPCR on the saliva samples and compared the results with the gold standard (a nasopharyngeal swab of the same patient). Saliva samples analyzed by RT-qPCR, reverse transcription isothermal amplification (RT-LAMP), spectroscopy, and enzyme-linked immunosorbent assay (ELISA) offer high sensitivity to detect SARS-CoV-2 in the early stages of the disease and among asymptomatic patients as compared to nasopharyngeal swab RT-qPCR. In addition, the self-collection of saliva offers the possibility of receiving telemedicine instructions to carry out the test, reducing the risk of contagion. The diagnosis of COVID-19 through saliva is sensitive, non-invasive, and is of low risk for the healthcare professionals. However, further studies are recommended to validate its clinical use.