Preclinical studies on liraglutide have suggested related improvements in β-cell function. Therefore, we investigated these effects in patients with type 2 diabetes (T2D) using the glucagon stimulation test (GST). We conducted a retrospective cohort study of 73 insulin-treated patients with T2D who had their treatment switched to liraglutide monotherapy. Their β-cell function was measured using a 1-mg intravenous GST at baseline and 24 weeks after treatment. The effect of liraglutide treatment on β-cell function was assessed by the change in the area under the curve (AUC) of serum C-peptide immunoreactivity during the GST (AUC-CPR). The AUC-CPR increased after 24 weeks of liraglutide treatment (9.80 ± 0.55 ng/mL⋅min to 11.50 ± 0.52 ng/mL⋅min, p = 0.001). In the univariate and adjusted multivariate regression analyses, a negative relationship between the change in the AUC-CPR and T2D duration was noted (β = -0.22, 95% confidence interval [CI] = -0.35 to -0.09, R² = 0.14, p = 0.001 and β = -0.20, 95% CI = -0.34 to -0.05, R² = 0.23, p = 0.008, respectively). In the analysis using T2D duration tertiles, early liraglutide treatment (T2D duration ≤10 years) significantly improved the AUC-CPR (<4 years: +2.56 ± 0.73 ng/mL⋅min, p = 0.002; 4-10 years: +2.60 ± 0.56 ng/mL⋅min, p < 0.001), whereas late liraglutide treatment did not (>10 years: -0.33 ± 1.15 ng/mL⋅min, p = 0.78). We conclude that early liraglutide treatment potentially improves β-cell function and subsequently glycemic control in patients with T2D, preventing further diabetic complications.