Objective To evaluate the application value of a novel fluorescent polymerase chain reaction (PCR)-melting curve method (referred to as "melting curve method") in detecting the resistance of Mycobacterium tuberculosis complex (MTBC) to isoniazid (INH) and fluoroquinolones (FQS). Methods Sputum specimens from hospitalized patients with pulmonary tuberculosis in Hunan Chest Hospital from April to October 2024 were collected prospectively. According to inclusion and exclusion criteria, 231 pathogenic positive specimens from tuberculosis patients were selected for Sanger sequencing, BACTEC MGIT 960 (referred to as "MGIT 960") liquid culture and phenotypic drug susceptibility testing, INH and FQS resistance-related gene mutations were detected with melting curve method, 28 specimens with negative melting curve results were excluded. A total of 203 MTBC-positive specimens were finally included in the analysis. Sanger sequencing and phenotype drug susceptibility results were taken as reference standards, respectively, the sensitivity, specificity, positive predictive value, and negative predictive value of the melting curve method for detecting INH and FQS resistance were calculated, consistency of the two was analyzed (kappa test). Results Taking Sanger sequencing as a reference standard, the sensitivity of the melting curve method in detecting MTBC resistance to INH and FQS was 97.92% (95%CI: 89.11%-99.63%) and 100% (95%CI: 88.97%-100%), respectively, the specificity was 100% (95%CI: 97.47%-100%) and 100% (95%CI: 97.72%-100%), respectively. The two methods showed high consistency (kappa values were 0.986 and 1.000, respectively). Taking phenotype drug susceptibility testing as a reference standard, the sensitivity of the melting curve method in detecting MTBC resistance to INH and FQS was 96.67% (95%CI: 83.33%-99.41%) and 94.44% (95%CI: 74.24%-99.01%), respectively, the specificity was 98.55% (95%CI: 92.24%-99.74%) and 97.53% (95%CI: 91.44%-99.32%), respectively. The two methods showed high consistency (kappa values were 0.952 and 0.900, respectively). Conclusion The melting curve method has high sensitivity and specificity in detecting INH and FQS drug-resistant mutations, which is of great value for the early and rapid diagnosis of drug-resistant tuberculosis.