Background <p>The automated insulin delivery (AID) systems are increasingly used in patients with type 2 diabetes requiring insulin therapy. This study aims to evaluate the effectiveness and safety of AID systems in type 2 diabetes.</p> Methods <p>Embase, PubMed, Web of Science, the Cochrane Library, and ClinicalTrials.gov website were searched until 25 May 2025. Randomized controlled trials (RCTs), single-arm trials, and real-world evidence (RWE) studies comparing AID with conventional insulin therapy in type 2 diabetes were included. Time in range (TIR, 3.9–10 mmol/L) was the primary outcome. Secondary outcomes included time below range (TBR, &lt; 3.9 mmol/L), time above range (TAR, &gt; 10 mmol/L), HbA_1c level, and safety outcomes.</p> Results <p>This meta-analysis included 10 RCTs (673 participants), 5 single-arm trials (460 participants), and 5 RWE studies (2147 participants). AID showed increased TIR in RCTs (mean differences [MD] 20.3%, 95% CI 14.86 to 25.73), single-arm trials (MD 15.55%, 95% CI 7.74 to 23.37), and RWE studies (MD 9.72%, 95% CI 7.58 to 11.86). The favourable effect of AID was also reflected in reductions in TAR (MD -15.97%, 95% CI -19.28 to -12.66) and HbA_1c (MD -0.79%, 95% CI -0.99 to -0.59), but not in TBR (MD -0.03%, 95% CI -0.08 to 0.02). Additionally, there was no much concern in serious adverse events.</p> Conclusions <p>AID systems are effective and safe for improving glycemic outcomes in insulin-requiring type 2 diabetes in both clinical trials and RWE studies.</p> Registration <p><i>This study was registered on PROSPERO (CRD</i>420251074781<i>)</i> <a href="https://www.crd.york.ac.uk/PROSPERO/view/CRD420251074781">https://www.crd.york.ac.uk/PROSPERO/view/CRD420251074781</a>.</p>