Two molecular epidemiologic methods, IS restriction fragment length polymorphism (IS-RFLP) and 24-locus mycobacterial interspersed repetitive unit-variable-number tandem repeat (MIRU-VNTR), are used worldwide in studies of (MTB). Conversely, because of its poor resolution, pulsed-field gel electrophoresis (PFGE) is not widely used for MTB. In this study, we improved the 24-locus MIRU-VNTR and PFGE protocols and compared the effectiveness of these approaches for the molecular typing of MTB using 75 clinical isolates obtained from a cohort investigation of high-risk populations infected with MTB. The 24-locus MIRU-VNTR method demonstrated superior discriminatory ability, followed by PFGE and IS-RFLP. Next, we analyzed six isolates with clear epidemiologic connections; that is, isolates from patients who attended the same school. IS-RFLP and PFGE identified these samples as the same type. By contrast, according to MIRU-VNTR, two isolates differed from four other isolates at one locus each; one isolate was identified as Mtub29 and the other as QUB-26. In summary, the 24-locus MIRU-VNTR assay was the most useful molecular typing method among the three methods investigated due to its discriminatory power, short time required, and availability as an epidemiologic investigation tool. PFGE was the second-best method. Compared with the other loci assessed in the 24-locus MIRU-VNTR assay, the Mtub29 and QUB-26 loci appeared to exhibit greater variability during transmission.