Objective To establish a new method for detecting Mycobacterium tuberculosis (TB) infection based on flow cytometry fluorescence immunoassay and interferon γ release assay (IGRA). Methods The whole blood samples were stimulated to produce interferon γ (IFN-γ) with phytagglutinin (PHA) and TB specific mixed peptides (early secretory antigenic target[ESAT-6] and culture filtrate protein 10[CFP-10]), and the plasma was analyzed by double antibody sandwich method combined with flow cytometry fluorescence immunoassay. The IFN-γ concentration was evaluated by receiver operating characteristic (ROC) curve for the diagnostic efficacy of TB. The linear range, minimum detection limit, repeatability, anti-interference performance of the established method were observed, and the consistency of detection with similar products on the market was evaluated. Results The linearity of the flow cytometry fluorescence immunoassay ranged from 2 pg/mL to 1 000 pg/mL. The lowest detection limit was 0.3 pg/mL; the repeatability parameters (coefficient of variation) of the samples at 100 pg/mL and 500 pg/mL were 4.58% and 2.46%, respectively. The average recovery rate of recovery assay was 98.0%. There was no interference with flow cytometry fluorescence immunoassay when the highest concentrations of triglyceride, bilirubin and hemoglobin were 50 mg/mL, 0.6 mg/mL and 10 mg/mL, respectively. As the optimum cut-off value of the IFN-γ concontration was 10 pg/mL, the sensitivity of IFN-γ in diagnosis of TB infection was 82.46% and the specificity was 87.30%. The total coincidence rates with T-SPOT, QFT, and Wantai TB-IGRA reagent were 97.2%, 83.0%, and 85.4%, respectively; and the Kappa coefficients were 0.822, 0.622 and 0.630, respectively. Conclusion The method for diagnosis of TB infection established in this study has a good performance, with the accuracy reaching the level of similar products on the market, and our method has obvious advantages in terms of repeatability and detection process.