The outbreak of severe acute respiratory syndrome (SARS) was caused by a newly identified coronavirus (SARS‐CoV) in 2003. To detect early SARS‐CoV infection, a one‐step, real‐time quantitative reverse transcription‐polymerase chain reaction (RT‐PCR) assay was developed that could simultaneously detect nucleocapsid (N), membrane (M), and spike (S) genes of SARS‐CoV with the same PCR condition using either Applied Biosystems (ABI) Prism 7700 Sequence Detection System or Roche LightCycler. The sensitivity of this assay was evaluated using cell culture‐derived viruses, in vitro transcribed viral RNA, and clinical specimens. The SARS‐S, ‐M, and ‐N primer/probe sets described in this paper could detect one to ten copies of in vitro transcribed S, M, and N RNA per test using both the ABI and Roche assay systems. The relative sensitivities for detecting cell culture‐derived SARS‐CoV were 0.01, 0.01, and 0.001 PFU/test, respectively. It showed that SARS‐N has comparable detection efficiencies to SARS2 and SARS3 which are primers sets designed by Centers for Disease Control and Prevention. In addition, SARS‐S and SARS‐M also demonstrated equivalent sensitivity to the commercially available RealArt HPA‐Coronavirus reagents (Artus). The relative sensitivity of these primer/probe sets was also examined using human sera spiked viruses and clinical specimens from four confirmed SARS patients. Similar results as above were obtained. Specificity tests and sequence alignment showed that these primer/probe sets annealed perfectly to 31 isolates of SARS‐CoV; and there was no cross detection with other coronaviruses and human respiratory tract‐associated viruses. Therefore, not only is it compatible with the ABI and Roche systems, this multiple‐gene detection assay also has the merit of being a rapid, safe, sensitive, and specific tool for accurate diagnosis of SARS‐CoV infection. J. Med. Virol. 77:151–158, 2005. © 2005 Wiley‐Liss, Inc.