Background/Objectives: Type 2 diabetes mellitus (T2DM) is characterized by progressive β-cell dysfunction and insulin resistance. Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) may enhance β-cell function. Semaglutide, a long-acting GLP-1 RA, improves glycemic control and weight, but its direct effects on β-cell function remain uncertain. Methods: This systematic review and meta-analysis followed PRISMA guidelines and was registered in PROSPERO (CRD420251034071). PubMed, Embase, and Scopus were searched through April 2025 for randomized controlled trials evaluating semaglutide’s effects on β-cell function in adults with T2DM. Primary outcomes included HOMA-B, HOMA-IR, and the proinsulin/insulin ratio; secondary outcomes included insulin secretion rate, insulinogenic index, and C-peptide. Two reviewers independently performed data extraction and risk-of-bias assessment using the Cochrane RoB 1 tool. Random-effects models were used for pooling. Certainty of evidence was evaluated using GRADE. Results: Sixteen trials (n = 6591) met inclusion criteria, with nine included in the meta-analysis. Semaglutide improved β-cell function (HOMA-B log ratio of means 1.50, 95% confidence interval [CI]: 1.25–1.80) and reduced insulin resistance (HOMA-IR ratio 0.82, 95% CI: 0.73–0.94) compared with placebo or active comparators. The pooled treatment ratio for proinsulin/insulin was 0.70 (95% CI: 0.63–0.79). However, risk of bias was generally high due to open-label designs, and certainty of evidence for all primary outcomes was rated very low. Conclusions: Semaglutide appears to improve β-cell function and insulin sensitivity in adults with T2DM, but conclusions remain uncertain given the very low certainty of evidence and substantial heterogeneity. High-quality trials with standardized β-cell outcomes are needed to confirm these findings.