A rapid, simple and highly discriminatory DNA fingerprinting methodology which produces data that can be easily interpreted, compared and transported is the ultimate goal for studying the epidemiology of Mycobacterium tuberculosis. A novel TaqI fluorescent amplified fragment length polymorphism (fAFLP) approach to M. tuberculosis DNA fingerprinting that targeted the variable IS6110 marker was developed in this study. The new method was tested for specificity and reproducibility, and compared with the standard reference IS6110 restriction fragment length polymorphism (RFLP) method for a panel of 78 isolates. Clustering conflicts between the two methods were resolved using mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTR) data. Comparison with an in-silico digestion of strain H37Rv showed that fAFLP-detected fragments were highly specific in vitro. The reproducibility of repeated digestions of strain H37Rv was 100%. Clustering results obtained by fAFLP and RFLP were highly congruent, with fAFLP allocating 97% of RFLP-clustered isolates to the same eight clusters as RFLP. Two single-copy isolates that had been clustered by RFLP were not clustered by fAFLP, but the MIRU-VNTR patterns of these isolates were different, indicating that the RFLP data had falsely clustered these isolates. Analysis by fAFLP will allow rapid screening of isolates to confirm or refute epidemiological links, and thereby provide insights into the frequency, conservation and consequences of specific transposition events.