The isolation of potentially pathogenic mycobacteria (PPM) from clinical specimens has become very frequent in the last years. Such organisms are typically environmental and occasionally pathogenic for humans. Standard diagnosis of mycobacterial infections relies on direct examination and culture. Nowadays, molecular tools are available, allowing quicker accurate diagnosis. Detection of PPM can be performed directly from clinical samples, although in most cases identification is carried out after isolation. Sequencing of genomic targets (such as 16S rRNA, rpoB or hsp65) allows accurate and quick identifications but has some technical limitations. Problems concerning sequencing analysis used for PPM identification together with description of available algorithms for PPM identification are the major objectives of this review.