Nasopharyngeal swabs (NPS) are generally used as specimen samples for antigen qualitative tests for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The principle of the reaction to the antigen protein is the same when saliva is used, and saliva samples were reported to be as accurate as NPS for real-time reverse transcription quantitative polymerase chain reaction (RT-qPCR) testing to identify SARS-CoV-2. Unlike NPS collection, self-collected saliva does not expose healthcare workers to the risk of infection. In this study, we evaluated the feasibility of using saliva samples for a SARS-CoV-2 antigen qualitative test (TA2107SA) under development. Saliva samples were collected from patients with confirmed or suspected COVID-19 infection and analyzed. The sensitivity, specificity, and concordance index of the antigen qualitative test were calculated using an RT-qPCR test as reference. Saliva samples were collected from 105 patients. The mean interval from onset to specimen collection was 5.7 days. The mean cycle threshold (Ct) value of RT-qPCR was 31.3. The sensitivity, specificity, and concordance index were 70.7%, 100%, and 0.85, respectively. In 33 patients with Ct values <30, the results of both the RT-qPCR and antigen tests were positive. The sensitivity of the saliva-based TA2107SA SARS-CoV-2 antigen qualitative test was slightly lower than that of the conventional antigen qualitative test using NPS samples from the same patient. Saliva-based antigen qualitative tests for SARS-CoV-2 are an alternative option during a pandemic.