We developed a PCR-based method to detect and quantify viable Bifidobacterium bifidum BF-1 cells in human feces. This method (PMA-qPCR) uses propidium monoazide (PMA) to distinguish viable from dead cells and quantitative PCR using a BF-1-specific primer set designed from the results of randomly amplified polymorphic DNA analysis. During long-term culture (10 days), the number of viable BF-1 cells detected by counting the number of CFU on modified MRS agar, by measuring the ATP contents converted to CFU, and by using PMA-qPCR decreased from about 10 10 to 10 6 cells/ml; in contrast, the total number of (viable and dead) BF-1 cells detected by counting 4=,6-diamidino-2-phenylindolee (DAPI)-stained cells and by using qPCR without PMA and reverse transcription-qPCR remained constant. The number of viable BF-1 cells in fecal samples detected by using PMA-qPCR was highly and significantly correlated with the number of viable BF-1 cells added to the fecal samples, within the range of 10 5.3 to 10 10.3 cells/g feces (wet weight) (r > 0.99, P < 0.001). After 12 healthy subjects ingested 10 10.3 to 10 11.0 CFU of BF-1 in a fermented milk product daily for 28 days, 10 4.5 ؎ 1.5 (mean ؎ standard deviation [SD]) BF-1 CFU/g was detected in fecal samples by using strain-specific selective agar; in contrast, 10 6.2 ؎ 0.4 viable BF-1 cells/g were detected by using PMA-qPCR, and a total of 10 7.6 ؎ 0.7 BF-1 cells/g were detected by using qPCR without PMA. Thus, the number of viable BF-1 cells detected by PMA-qPCR was about 50 times higher (P < 0.01) than that detected by the culture-dependent method. We conclude that strain-specific PMA-qPCR can be used to quickly and accurately evaluate viable BF-1 in feces.