Abstract: Objective To establish a rapid method for detecting drug resistance of Mycobacterium tuberculosis （MTB） by using an indicator of MPT64 and to evaluate the application of MPT64 in detecting rifampicin （RFP） and isoniazid （INH） resistance of MTB. Methods The expression levels of MPT64 in the Middlebrook 7H9 liquid medium containing （observation group） and non-containing （control group） RFP and INH, separately cultured by MTB sensitive strains and multidrug-resistant strains for 3, 7, 10 days were detected by using gold immunotomographic assay （GICA）. The grayscale of the experimental strip was analyzed by photographs with white light source of gel imaging instrument （Image Lab Software 3.0）. The grayscale ratio was compared between the sensitive and multidrug-resistant strains. The cutoff value of the grayscale ratio for RFP and INH resistance was determined based on the analysis of the area under the receiver operating characteristic （ROC） curve （AUC）. The accuracy, sensitivity, specificity, positive and negative prediction values of this new method were evaluated with the proportional test results as the gold standard. Results With the prolongation of culture time, the grayscale ratios of 12 MTB sensitive strains and 11 multidrug-resistant strains were increased in the control group. The grayscale ratios of sensitive strains at different culture times did not change significantly in the observation group, whereas those of the multidrug-resistant strains were significantly increased. The MPT64 grayscale ratios of sensitive and multidrug-resistant strains in the drug medium significantly differed at 3, 7 and 10 days （all P < 0.05）. According to the grayscale ratio in the observation group, the ROC curves of the RFP and INH resistance detection were plotted with the indicator of MPT64. On day 3, the AUCs of RFP and INH were 0.84 and 0.77, and the grayscale ratios were both 0.05. On day 7 and day 10, the AUCs were both 1.00. On day 7, the grayscale ratios were 0.23 and 0.20. On day 10, the grayscale ratios were 0.49 and 0.48. The accuracy of GICA in detecting RFP and INH drug-resistance of 34 MTB was 97% and 94%, 93% and 86% for the sensitivity, 100% and 100% for the specificity, 100% and 100% for the positive prediction value, and 95% and 91% for the negative prediction value, respectively. Conclusions MPT64 detection by GICA can accurately and rapidly identify the RFP and INH resistance on day 7 of MTB culture. MPT64 can be used as an effective detection indicator for MTB drug sensitivity.