The T-SPOT.TB test does not cross-react with Bacille Calmette-Guérin or most non-tuberculosis mycobacterium species, and is based on IFN-γ responses to Mycobacterium tuberculosis-specific antigens. The objective of this study was to compare tuberculin skin test (TST) with T-SPOT.TB results used in the diagnosis of active tuberculosis (TB) as well as latent tuberculosis infection (LTBI). A total of 136 subjects participated in three different groups (47 patients with active pulmonary TB, 47 healthy persons without M. tuberculosis exposure, and 42 hospital members with a history of close contact with active TB patients). The T-SPOT.TB sensitivity (83.0%) and the negative predictive value (NPV) (82.6%) in the diagnosis of active TB were significantly higher than those of TST. The sensitivity and NPV of the TST were 38.3 and 60.8%, respectively. The T-SPOT.TB specificity (80.9%) and positive predictive value (81.3%) were lower than those of TST (95.7 and 90.0%, respectively). The performance of T-SPOT.TB and TST for diagnosing LTBI was the same (54.8%). T-SPOT.TB was superior in terms of sensitivity (83.0%); TST detected only 18, whereas T-SPOT.TB test detected 39 out of 47 patients with active TB. T-SPOT.TB is thought to have better performance than TST due to false-negative results in diagnosing active TB. However, it is considered that large prospective longitudinal studies are needed for diagnosing LTBI.