Omicron is the current variant of SARS-CoV-2. It has high transmissibility and evades human immunity. Early and accurate diagnosis is essential for therapy, prognosis, and curbing the spread of the virus. Here we reported a new sampling method for the rapid antigen test, with specimens from phlegm, instead of from nasal swabs. Sensitivities in detection of SARS-CoV-2 antigen were compared within the two specimens, and between rapid antigen tests with phlegm specimens and PCR tests. Of 41 volunteers, thirty-one with positive phlegm specimens of SARS-CoV-2 eventually had typical COVID-19 symptoms, which suggested rapid antigen tests with phlegm specimens had 100 % accuracy. Fourteen of these had antigen tests with nasal swab specimens: 13 negative and 1 positive. Combined with that the antigen levels of the 6561-fold diluted phlegm specimen were comparable to those of the original nasal swab specimen, antigen tests with phlegm specimens are more sensitive and earlier detect SARS-CoV-2 than with nasal swab specimens. Interestingly, case studies indicated antigen tests with phlegm specimens earlier notified patients of positive infection than PCR tests. Phlegm specimens enhanced sensitivity in detection of SARS-CoV-2 in antigen tests, resulting in earlier diagnosis (12 to 42 hours, n=6) than nasal swab specimens. The sensitivity of antigen tests with phlegm specimens is comparable to that of PCR tests, but the former earlier outputs test results. Rapid antigen tests with phlegm specimens facilitate monitoring the health of COVID-19 patients and direct recovery.