Transplantation of insulin-producing cells derived from pluripotent stem cells represents a highly promising approach for the radical treatment of type 1 diabetes (T1D). Informed by a comprehensive understanding of fetal pancreatic development, directed differentiation protocol for generating pancreatic β cells from pluripotent stem cells has been established and has achieved considerable advances, enabling the production of mature, fully functional β cells that closely recapitulate the characteristics of native pancreatic β cells. Preclinical studies have shown that the transplantation of stem cell-derived islets (SC-islets) reverses hyperglycemia in both mouse and nonhuman primate models, with a favorable safety profile. Early-phase clinical trials have further corroborated the safety and efficacy of this approach, a subset of patients with long-standing T1D achieved insulin independence, described as a "functional cure", with no serious adverse events of clinical significance reported. Despite these encouraging results, substantial challenges remain. With respect to differentiation protocols, insufficient functional maturity, pronounced cellular heterogeneity, significant batch-to-batch variability, and the challenges of large-scale manufacturing represent the principal unresolved limitations. Of particular concern, immune rejection remains a critical barrier even after the transplantation of autologous SC-islets, necessitating continued reliance on immunosuppressive therapy. Cell encapsulation and gene editing strategies have emerged as potential approaches to overcome this immunological barrier. In this review, we discuss strategies for obtaining insulin-producing cells from diverse cellular sources, summarize the latest advances in stem cell-based diabetes therapy, and propose future research directions.