The purpose of the article is to compare the whole blood interferon-γ release assay (IGRA) with the traditional methods for detecting Mycobacterium tuberculosis (MTB) infection in children. Fifteen childhood patients with tuberculosis and 15 healthy children were recruited. Sputa samples and venous blood were collected, and according to different procedures, IGRA, sputum smear, colloidal gold assay (CGA), fluorescence quantitation polymerase chain reaction (FQ-PCR), and tuberculosis skin test (TST) were, respectively, performed. Thirty healthy children vaccinated with Bacillus Calmette-Guérin (BCG) were also recruited, and the comparative test was carried out between IGRA and TST. In all of 15 childhood patients with TB, the positive rates were 86.7%, 20.0%, 26.7%, 40%, and 66.7% in IGRA, sputum smear, CGA, FQ-PCR, and TST, respectively. In the children vaccinated with BCG, the positive rate of IGRA was significantly lower than that of TST (6.7% vs 76.7%). From high to low, the specificities of the five methods were sputum smear (100%), IGRA (86.7%), FQ-PCR (86.7%), TST (40%), and CGA (26.7%). Although the specificities of sputum smear and FQ-PCR were more than or equal to that of IGRA, the relative sensitivities limited their applications in populations of children. IGRA is a sensitive and specific method, and could be taken as a first choice for detecting MTB infection in populations of children.