Abstract: This study aimed to simultaneously enhance the enzymatic activity and thermostability of glucose oxidase (AnGOD) derived from Aspergillus niger through a semi-rational design strategy. Mutation sites were designed using the FireProt online tool and consensus sequences analysis, and the optimal mutant was screened via site-directed mutagenesis and a two-stage combinatorial mutation approach. Molecular dynamics simulations were then conducted to elucidate the mechanism of performance enhancement. Results showed that from the initial 83 designed sites, 12 single-point advantageous mutants were selected, and the optimal mutant M5 (T276F/T34V/Q90R/S53F/S74T) was obtained through two-stage combinatorial mutagenesis. Compared to the wild-type (WT), M5 exhibited a 4.2-fold increase in specific activity, with a k_cat value of 755.0 s⁻¹ and a k_cat/K_m value of 34.52 mmol⁻¹s⁻¹, representing 3.67- and 2.48-fold enhancements over WT, respectively. At 65 ℃, the half-life of M5 was 30 min, representing a 30-fold increase compared to WT (1 min). Molecular dynamics simulations revealed that the average RMSD of M5 and the RMSF in non-catalytic regions were reduced by 12.9% and 12.4%, respectively, compared to the wild-type (WT). Additionally, the number of hydrogen bonds between M5 and the substrate/coenzyme increased by 10.3% and 8.8%, respectively. Notably, the S74T mutation introduced a novel hydrogen bond network, while the Q90R mutation introduced a positive charge, enhancing surface electrostatic complementarity. These synergistic effects collectively improved both the thermostability and catalytic activity of AnGOD. Through semi-rational design and combinatorial mutagenesis, this study significantly alleviated the stability-activity trade-off in AnGOD enzyme engineering, laying a foundation for the application of GOD.