In this study, we determined the MICs and MBCs of levofloxacin (LVFX), clarithromycin (CAM), and KRM-1648 (KRM) for Mycobacterium tuberculosis (MTB) strain Kurono and M. avium complex (MAC) strain N-444 residing in MONO-MAC-6 human macrophage-like cells (MM6-M phi s) and A-549 human type II alveolar epithelial cells (A-549 cells). First, the MICs of LVFX for MTB replicating in MM6-M phi s (1 microgram/ml) and A-549 cells (2 micrograms/ml) were 4 to 8 times higher than its MICs for extracellular MTB growing in 7HSF medium. In contrast, the MICs of CAM for intracellular MTB residing in both the cells (2-4 micrograms/ml) were 4 to 8 times less than its MICs for extracellular MTB organisms. On the other hand, the MICs of KRM for extracellular MTB were nearly the same as its MICs for intracellular MTB residing in both types of the cells. Secondly, the MICs of LVFX and CAM for extracellular MAC were not significantly different from their MICs for intracellular MAC residing in MM6-M phi s and A-549 cells. The MIC of KRM for MAC residing in A-549 cells was 0.25 microgram/ml, and this value was 32 times higher than its MIC for MAC residing in MM6-M phi s (0.008 microgram/ml). Thirdly, the MBCs of test drugs for intracellular MTB and MAC residing in both types of the cells were somewhat longer than their MBCs for extracellular organisms. These findings indicate that, in the case of pulmonary infections with MTB or MAC, the therapeutic efficacy of a given drug, especially KRM, is more or less influenced by the bacterial location in the host lung tissues where the mycobacterial pathogens survive and multiply, i.e., either alveolar M phi s, type II alveolar epithelial cells, or surrounding environment.