Antemortem detection of Mycoplasma hyopneumoniae infection in swine production systems has relied on antibody testing, but the availability of tests based on DNA detection and novel diagnostic specimens, e.g., tracheal swabs and oral fluids, has the potential to improve M. hyopneumoniae surveillance. A field study was performed over a 14-week period during which 10 pigs in one pen at the center of a room with 1,250 6-week-old pigs housed in 46 pens were intratracheally inoculated with M. hyopneumoniae Thereafter, one tracheal sample, four serum samples, and one oral fluid sample were collected from every pen at 2-week intervals. Tracheal and oral fluid samples were tested for M. hyopneumoniae DNA and serum samples for M. hyopneumoniae antibody. Test results were modeled using a hierarchical Bayesian model, based on a latent spatial piecewise exponential survival model, to estimate the probability of detection by within-pen prevalence, number of positive pens in the barn, sample allocation, sample size, and sample type over time. Analysis showed that tracheal samples provided the earliest detection, especially at large sample sizes. While serum samples are more commonly collected and are less expensive to test, high probability of detection estimates were only obtained 30 days postexposure at large sample sizes. In all scenarios, probability of detection estimates for oral fluids within 30 days were significantly lower than those for tracheal and serum samples. Ultimately, the choice of specimen type, sample number, and assay will depend on testing objectives and economics, but the estimates provided here will assist in the design of M. hyopneumoniae surveillance and monitoring programs for different situations.