Objective To evaluate the detection performances of two interferon-gamma (IFN-γ）release assay (IGRA) kits for mycobacterium tuberculosis (MTB) infection. Methods One hundred and five clinical blood samples were detected (61 in tuberculosis group, 31 in nontuberculous pulmonary disease group, and 13 in healthy control group), 82 of which were detected with two kits contrastively (43 in tuberculosis group, 26 in nontuberculous pulmonary disease group, and 13 in healthy control group), and QuantiFERON-TB Gold (Cellestis, Australia) was served as reference kit to calculate the accuracy of test kit in detection of infection and to explore the basic immune status by counting the quantity of IFN-γ in tuberculosis group and healthy control group. Results The sensitivity, specificity, positive predictive value, negative predictive value and diagnostic validity of reference kit were 81.39% (35/43), 76.92% (30/39), 79.55% (35/44), 78.95% (30/38) and 79.27% (65/82) respectively, and those of test kit were 73.77% (45/61), 84.09% (37/44), 86.54% (45/52), 69.81% (37/53) and 78.09% (82/105) respectively. The coincidence rate of two kits was 86.56% (Kappa=0.734). The quantity of IFN-γ after stimulation in tuberculosis group was significantly higher than that in healthy control group (t=2.529, P=0.014). Conclusion The detection performances of test kit and reference kit are almost the same, and both kits can work for the auxiliary diagnosis of MTB infection. T cell activity may be increased reactively to some extent in patients with tuberculosis.