Background: Self-collected saliva samples can increase the diagnostic efficiency and benefit healthcare workers, patient care, and infection control. This study evaluated the performance of self-collected saliva samples compared to nasopharyngeal swabs using three commercial kits for the qualitative detection of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Methods: Matched nasopharyngeal and saliva samples were collected from 103 patients with either asymptomatic or symptomatic COVID-19. Both samples were evaluated using three commercial kits (TaqCheck, Allplex, and TaqPath). To evaluate sample stability, viral RNA extraction was performed in the presence or absence of an RNA-stabilizing solution. Storage conditions, including the duration, temperature, and stability after freezing and thawing of the samples, were also evaluated. Results: All the saliva samples showed 100% concordance with the nasopharyngeal swab results using TaqCheck and Allplex kits, and 93% using TaqPath kit. No difference was observed in the samples that used the RNA-stabilizing solution compared to the group without the solution. The Ct values of the freeze–thawed samples after 30 days were higher than those on day 0; however, the results were consistent the fresh samples. Conclusion: The high concordance of SARS-CoV-2 detection via reverse transcription–polymerase chain reaction (RT-PCR) in matched saliva and nasopharyngeal samples using different commercial assays reinforces the concept that self-collected saliva samples are non-invasive, rapid, and reliable for diagnosing SARS-CoV-2 infection.