BACKGROUND: This study aimed to test the efficacy of patient-centered self-management intervention (PACE-SMI) to improve HbA1c, self-efficacy, and self-care behaviors in adults with type 2 diabetes mellitus (T2DM). METHODS: In this multicenter, parallel two-arm randomized controlled trial, 612 adults with T2DM and HbA1c ≥ 7% were enrolled and assigned to the control group (n = 310) and the intervention group (n = 302) using stratified permuted block randomization. The control group received usual care, whereas the intervention group received usual care plus nurse-led, theory-driven, culturally tailored PACE-SMI, comprising eight weekly sessions of individualized education, counseling, behavioral training, and home visit. Outcomes were assessed at baseline, postintervention, and 3 months follow-up. RESULTS: Data at 3 months were provided by 583 participants (control: n = 295, intervention: n = 288). Per-protocol analysis showed that the intervention group had a lower mean HbA1c (8.49% [standard deviation (SD), 1.58]) than the control group (8.74% [SD, 1.62]), with small yet statistically significant mean difference of 0.25% (95% confidence interval [CI], -0.01 to 0.51; Cohens d = 0.16; p = 0.03). Self-efficacy and self-care behaviors significantly improved in the intervention group (116.89 [SD, 25.50] and 70.01 [SD, 17.97]) compared to the control group (75.43 [SD, 18.99] and 51.54 [SD, 12.04]), with mean differences of 41.48 (95% CI, 37.83-45.13; Cohens d = 1.84; p < 0.0001) and 18.56 (95% CI, 16.08-21.04; Cohens d = 1.22; p < 0.0001), respectively. Linear regression analysis indicated the effect of PACE-SMI on HbA1c was significantly mediated by improvements in self-efficacy and self-care behaviors (R2 = 0.232, p < 0.001). CONCLUSION: PACE-SMI led to modest but significant improvement in HbA1c and substantial enhancements in self-efficacy and self-care behaviors in adults with T2DM.