The frequency of false-positive cultures for Mycobacterium tuberculosis due to cross-contamination has been difficult to determine because of the lack of specific strain markers. Isolates collected prospectively over 5 yr from a municipal health department laboratory underwent DNA fingerprinting using the IS6110 and pTBN12 sequences. We reviewed the clinical and laboratory records of all isolates that had matching DNA fingerprints and were processed within 42 d of each other; 8 isolates were classified as probable or definite false-positives, representing 4.0% (8/199) of the culture-positive patients. A convenience sample of 42 isolates from three other mycobacterial laboratories also underwent DNA fingerprinting, and five (12%) were found to be definite or probable false-positives. Cross-contamination during initial processing of specimens was the most common source of false-positive cultures. The source of cross-contamination for three false-positive cultures was a laboratory proficiency survey specimen containing strain H37Ra. Ten of the 13 patients were misdiagnosed as having tuberculosis, and seven received unnecessary multidrug treatment. Clinicians should be aware of the potential for false-positive cultures for M. tuberculosis, and mycobacteriology laboratories need to carefully review procedures to minimize this occurrence. DNA fingerprinting provides a valuable tool for the study of false-positive cultures.