A risk analysis of Mycobacterium avium subsp. paratuberculosis (MAP) in milk harvested using good milking practice and treated in validated pasteurisers was carried out. We used the results from a survey of MAP in New Zealand raw milk silos in conjunction with data from our earlier kinetic study of the survival of MAP after high temperature–short time (HTST) pasteurisation, as well as data from two other published studies. Raw milk samples were collected at approximately weekly intervals over a 12-month period from raw milk silos at six dairy plants located throughout the major New Zealand dairying areas. One hundred and seventy five raw milk samples were forwarded chilled by overnight delivery to an accredited testing laboratory (AgResearch) for culturing. Over that period these samples represented a total volume of 4.6 million cubic metres of raw milk. One 50 ml aliquot was confirmed as being MAP culture-positive. A first-order kinetic heat inactivation model was developed to predict the likelihood of MAP being present in milk immediately after pasteurisation. The model took into account the expected concentration of MAP in the incoming raw milk supply (with recognition of the limitations imposed by the data set) and the variability of the lethality of the heat treatment on the five different strains reported in our earlier kinetic study. This variability was consistent with the variation between three strains reported by others from laboratory pasteurisation. The model simulations were run using @Risk software. The risk analysis demonstrated that good milking practice combined with properly operated and validated pasteurisers operating at 72°C/15 s will give an extremely low probability of any viable MAP surviving per cubic metre in pasteurised milk. As such, commercial pasteurisation can be considered to be an effective treatment step for the destruction of MAP.