beta-Glucosidase (BGL) plays a key role in cellulose hydrolysis. However, it is still a great challenge to enhance product tolerance and enzyme activity of BGL simultaneously. Here, we utilized one round error-prone PCR to engineer the Penicillium oxalicum 16 BGL (16BGL) for improving the cellulosic ethanol yield. We identified a new variant (L-6C), a triple mutant (M280T/V484L/D589E), with enhanced catalytic efficiency (k(cat)/K-m) for hydrolyzing pNPG and reduced strength of inhibition (K-m(app)/K-I) by glucose. To be specific, L-6C achieved a K-m(app)/K-I of 0.35 at a glucose concentration of 20 mM, which was 3.63 times lower than that attained by 16BGL. The catalytic efficiency for L-6C to hydrolyze pNPG was determined to be 983.68 mM(-1) s(-1), which was 22% higher than that for 16BGL. However, experiments showed that L-6C had reduced binding affinity (2.88 mM) to pNGP compared with 16BGL (1.69 mM). L-6C produced 6.15 g/L ethanol whose yield increased by about 10% than 16BGL. We performed molecular docking and molecular dynamics (MD) simulation, and binding free energy calculation using the Molecular Mechanics/Poisson Boltzmann surface area (MM/PBSA) method. MD simulation together with the MM/PBSA calculation suggested that L-6C had reduced binding free energy to pNPG, which was consistent with the experimental data.