Heteroresistance in Mycobacterium tuberculosis isolates remains the major challenge for phenotypic drug susceptibility testing (DST) methods to detect drug resistance. The aim of this study was to investigate the abilities of phenotypic DST methods to identify the isoniazid (INH) heteroresistance in M. tuberculosis. We found that the broth dilution method was able to detect INH resistance if 0.5% resistant bacteria with mutations in the katG and oxyR-ahpC regions were present, while the detection limit ranged from 1 to 10% for the INH-resistant strains harboring inhA mutations, which was associated with the different mutant types. Additionally, MGIT DST was able to find the recommended 1% INH resistance due to katG mutations. In contrast, MGIT DST detected resistance in suspensions with 20% resistant bacteria with inhA mutations. Statistical analysis revealed that the ability of the broth dilution method to detect heteroresistance was better than that of the MGIT DST (p = 0.004). When we further pairwise compared the two methods for detecting heteroresistance according to different mutant loci, the broth dilution method found more heteroresistance due to inhA mutations than MGIT DST (p = 0.001), while the differences for katG and oxyR-ahpC mutations were both not statistically significant (p > 0.05). In conclusion, our findings demonstrate that MGIT DST fails to detect INH heteroresistance in M. tuberculosis isolates with mutations in the promoter region of inhA. In addition, the broth dilution method is more sensitive than MGIT DST in finding INH heteroresistance, indicating that this method may serve as an alternative method to detect the heteroresistance of M. tuberculosis.