Abstract Background Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) is the cause of Coronavirus Disease 2019 (COVID-19). Although Real Time Reverse Transcription Polymerase Chain Reaction (qRT-PCR) of respiratory specimens is the gold standard test for detection of SARS-CoV-2 infection, collecting nasopharyngeal swabs causes discomfort to patients and may represent considerable risk for healthcare workers. The use of saliva as a diagnostic sample has several advantages. Objectives The aim of this study was to validate the use of saliva as a biological sample for diagnosis of COVID-19. Methods This study was conducted at Infectious Diseases Research Laboratory (LAPI), in Salvador, Brazil. Participants presenting with signs/symptoms suggesting SARS-CoV-2 infection underwent a nasopharyngeal swab (NPS) and/or oropharyngeal swab (OPS), and saliva collection. Saliva samples were diluted in PBS, followed by RNA isolation and RT-Real Time PCR for SARS-CoV-2. Results of conventional vs saliva samples testing were compared. Statistical analyses were performed using Statistical Package for the Social Sciences software (SPSS) version 18.0. Results One hundred fifty-five participants were recruited and samples pairs of NPS/OPS and saliva were collected. The sensitivity and specificity of RT-PCR using saliva samples were 94.4% (95% CI 86.4-97.8) and 97.62% (95% CI 91.7-99.3), respectively. There was an overall high agreement (96.1%) between the two tests. Conclusions Use of self-collected saliva samples is an easy, convenient, and low-cost alternative to conventional NP swab-based molecular tests. These results may allow a broader use of molecular tests for management of COVID19 pandemic, especially in resources-limited settings.