Rapid diagnosis of mycobacterial infections is of crucial importance. For this reason molecular methods have become valuable diagnostic tools. DNA isolation step prior to polymerase chain reaction (PCR) is one of the most important steps that affect the sensitivity of the diagnostic tests. In this study, three isolation methods which could affect the sensitivity of the PCR methods used for the diagnosis of mycobacterial infections were compared. For this reason the sensitivities of boiling method, MagNA Pure LC automated isolation system (Roche Applied Science, Germany) and Magtration 12GC (Precision System Science, Germany) automated isolation system were compared by using four different mycobacterial strains (Mycobacterium tuberculosis H37Ra standard strain, M.tuberculosis clinical isolate, M.phlei and M.smegmatis). DNAs were isolated from serial dilutions of these mycobacterial strains by using three isolation methods, and 441 base pair region of hsp65 gene was amplified by PCR. After the obtained products were run on agarose gel, it was observed that the lowest dilution rates at which bands could be seen were 10(-5) dilutions of H37Ra DNA isolated by MagNA Pure and Magtration systems. Isolation of H37Ra DNA by boiling method yielded a band in 10(-2) dilution sample. Isolation of DNA from Mycobacterium, tuberculosis clinical isolate by MagNA Pure system, Magtration system and boiling method provided amplification in 10(-4), 10(-3) and 10(-1) dilution samples, respectively. The lowest dilution samples which yielded visible bands on agarose gel were 10(-5), 10(-4) and 10(-2) dilution samples for M.smegmatis strain and 10(-4), 10(-3) and 10(-2) dilution samples for M.phlei strain. The results of this study revealed that MagNA Pure system could detect smaller amounts of DNA in the sample while boiling method was not a reliable method to detect small amounts of DNA in clinical samples. Therefore it was concluded that the use of automated isolation systems could be a reliable alternative for routine PCR applications by providing both high sensitivity and standardization.