In this study, a new test based on a novel faecal nucleic acid extraction method and an IS900-based real-time quantitative PCR (QPCR) method was developed and evaluated for detection and quantification of Mycobacterium avium subsp. paratuberculosis(MAP) DNA in ovine and bovine faecal samples. Both the Cattle (C) and Sheep (S) strains of MAP were detected by the QPCR assay, and no cross reactions were detected with 51 other species of mycobacteria including 10 which contained IS900-like sequences. One copy of IS900 fragment cloned into plasmid pCR2.1 and 1 fg of MAP genomic DNA were consistently detected, while in spiked faecal samples the detection limit was 10 viable MAP (S strain) per one gram of negative ovine faeces. A total of 506 individual ovine faecal samples with known culture (BACTEC) results and histological status were tested in Australia, and a total of 666 bovine faecal samples cultured by Herrold's egg yolk medium (HEYM) were tested in Japan. In ovine samples, the QPCR assay detected 68 of 69 (98.6%) BACTEC culture positive faeces and there was a strong relation between time to detection in culture and DNA quantity measured by QPCR (r=-0.70). Furthermore, when DNA quantities detected by the QPCR were analysed on the basis of histological classification, faecal samples representing sheep with multibacillary lesions showed significantly higher levels of MAP DNA than samples from sheep with paucibacillary lesions or no lesions, suggesting that the QPCR test could be used for estimation of the risk of transmission. In bovine samples, 54 of 60 (90.0%) HEYM culture positive faeces were detected by the QPCR. MAP DNA was also detected from some culture negative faecal samples from sheep and cattle exposed to MAP, suggesting that the QPCR has very high analytical sensitivity for MAP in faecal samples and detects non-viable MAP in faeces. None of the faecal samples from 176 sheep and 508 cattle that were not exposed to MAP were positive in QPCR.