Tuberculosis (TB) represents a barrier for free trade of livestock between Mexico and the United States of America (US). In spite of efforts from Mexico to export TB-free animals, some of those found with TB lesions in slaughterhouses in the US are traced back to that country. Therefore, the purpose of this study was to determine, through molecular epidemiology, the most probable source of infection for cattle found with TB lesions in the US. Ninety M. bovis isolates, 50 from Mexico obtained from cattle in 8 different states, and 40 from the US from cattle, deer, elk and feral pigs from 7 different states were included in the study. All samples were analyzed in both laboratories, Mexico and the US, following the same protocol for molecular analysis by spoligotyping. Twenty-seven clusters, ranging from 1 to 18 genetically similar strains were found. Some clustering by country was observed, strains from cattle and deer in Michigan in the US fell into the same cluster, suggesting transmission between species. These results, combined with epidemiological information suggest that despite of the possibility that some animals with lesions in the US come from Mexico as false negatives, the US has its own source of infection, must probably in dairy cattle and wildlife. Genetic diversity of isolates from Mexico was larger than that in the US, which could be a consequence of the endemic status of the disease and the indiscriminate movement of animals between regions.