Objective: This meta-analysis aimed to compare the efficacy and safety of ultrasound-guided (US-guided) versus landmark-guided (LM-guided) local corticosteroid injection for carpal tunnel syndrome (CTS). Methods: Database including Pubmed, Embase, and Cochrane Library were searched to identify relevant randomized controlled trials (RCTs). The outcomes mainly included Boston Carpal Tunnel Questionnaire (BCTQ): Symptom Severity Scale (BCTQs), Functional Status Scale (BCTQf); and electrophysiological indexes: distal motor latency (DML), sensory distal latency (SDL), compound muscle action potential (CAMP), sensory nerve action potential amplitude (SNAP), and sensory nerve conduction velocity (SNCV). Adverse events were also recorded. Results: Overall, nine RCTs were finally screened out with 469 patients (596 injected hands). Pooled analysis showed that US-guided injection was more effective in BCTQs (SMD, −0.69; 95% CI, −1.08 to −0.31; P = 0.0005), BCTQf (SMD, −0.23; 95% CI, −0.39 to −0.07; P = 0.005), CAMP (MD, 0.64; 95% CI, 0.35−0.94; P < 0.0001) improvement, and a lower rate of adverse events (RR, 0.34; 95% CI, 0.22−0.52; P < 0.00001). Subgroup analysis revealed that the US-guided injection had significantly better CMAP than the LM-guided for the in-plane approach (MD, 0.69; 95% CI, 0.36−1.01; P < 0.0001) but not for the out-plane approach (MD, 0.39; 95% CI, −0.39 to 1.17; P = 0.33). Conclusions: US-guided injection was superior to LM-guided injection in symptom severity, functional status, electrodiagnostic, and adverse events improvement for CTS. To some extent, the in-plane approach yields better results compared with the out-plane process under US guidance.