Extensive studies have shown that the current assays used to identify cattle infected with Mycobacterium bovis or Mycobacterium avium subsp. paratuberculosis are not sufficiently sensitive and specific to detect all infected animals, especially animals recently infected with the pathogens. In the present report we show that these limitations might be overcome with a latex bead agglutination assay (LBAA). With the specific immunodominant epitope (ESAT6-p) of M. bovis, we developed an LBAA and enzyme immunoassay (EIA) for that purpose and compared them with the "gold standard" culture method and skin test for their efficacy in detecting bovine tuberculosis. When sera from control healthy cows (n = 10), M. avium subsp. paratuberculosis-positive cattle (naturally infected, n = 16; experimentally infected, n = 8), and M. bovis-positive cattle (naturally infected, n = 49;experimentally infected, n = 20) were applied to an EIA and an LBAA developed with ESAT6-p, the two tests showed similar sensitivity (97.1% by EIA, 95.7% by LBAA), high specificity (94.2% by EIA, 100% by LBAA), and a positive correlation (kappa value, 0.85; correlation rate, 93.2%; correlation coefficient, 0.64). Receiver operating characteristic analysis of EIA results and comparison with the culture method determined a suitable cutoff value at 0.469, with an area under the curve of 0.991 (95% confidence interval, 0.977 to 1.0). As LBAA didn't show any positive reactions with sera from uninfected control cows or M. avium subsp. paratuberculosis-infected cattle, which were confirmed to be free of M. bovis by culture or PCR, LBAA using the ESAT6-p can be a rapid and useful M. bovis diagnostic assay. The data suggest that rapid, sensitive, and specific assays can be developed with peptides containing immunodominant epitopes present in proteins uniquely expressed in M. bovis or M. avium subsp. paratuberculosis for differential diagnosis of cattle infected with M. bovis or M. avium subsp. paratuberculosis.