Mycobacterium bovis, a member of the Mycobacterium tuberculosis complex, is the most common agent of cattle tuberculosis, a zoonosis that causes losses in meat and milk production in several countries. In order to support epidemiological studies aimed at controlling the disease, several methods for molecular discrimination of M. bovis isolates have recently been developed. The most frequently used are spacer oligonucleotide typing (spoligotyping), mycobacterial interspersed repetitive units (MIRU), and exact tandem repeat (ETR), but they all have different discriminatory power. In the present study, allelic diversity was calculated for each MIRU and ETR locus, and the Hunter-Gaston discriminatory index (HGI) was calculated for spoligotyping, 10 MIRUs, and 3 ETRs, in 116 isolates of M. bovis obtained from cattle. The analysis of allelic diversity indicated that MIRUs 16, 26, and 27, and ETRs A, B, and C, showed the greatest diversity between the assayed loci. The HGIs for each of the techniques were: spoligotyping=0.738381; MIRU=0.829835; and ETR=0.825337. The associations of the methods' improved discriminatory power were: spoligotyping+MIRU=0.930585; spoligotyping+ETR=0.931034; and MIRU+ETR=0.953373. The greatest discriminatory power was obtained when the three techniques were associated (HGI=0.98051). Considering the analyses of the present study, spoligotyping should be the first method to be used because it differentiates M. bovis from the other members of the Mycobacterium tuberculosis complex. As the associations of MIRU and ETR with spoligotyping resulted in nearly identical HGIs, ETR seems to be the best choice after spoligotyping, because it is faster and more economical than MIRU. Finally, MIRU should be the last method used. In spite of this finding, the choice of the method used should be based on the discriminatory power necessary for the objective at hand.