Aim To compare the safety and effects of unrestricted visiting policies (UVPs) and restricted visiting policies (RVPs) in intensive care units (ICUs) with respect to outcomes related to delirium, infection, and mortality. Methods MEDLINE, Cochrane Library, Embase, Web of Science, CINAHL, CBMdisc, CNKI, Wanfang, and VIP database records generated from their inception to 22 January 2022 were searched. Randomized controlled trials and quasi-experimental studies were included. The main outcomes investigated were delirium, ICU-acquired infection, ICU mortality, and length of ICU stay. Two reviewers independently screened studies, extracted data, and assessed risks of bias. Random-effects and fixed-effects meta-analyses were conducted to obtain pooled estimates, due to heterogeneity. Meta-analyses were performed using RevMan 5.3 software. The results were analyzed using odds ratios (ORs), 95% confidence intervals (CIs), and standardized mean differences (SMDs). Results Eleven studies including a total of 3741 patients that compared UVPs and RVPs in ICUs were included in the analyses. Random effects modeling indicated that UVPs were associated with a reduced incidence of delirium (OR = 0.4, 95% CI 0.25-0.63, I-2 = 71%, p = 0.0005). Fixed-effects modeling indicated that UVPs did not increase the incidences of ICU-acquired infections, including ventilator-associated pneumonia (OR = 0.96, 95% CI 0.71-1.30, I-2 = 0%, p = 0.49), catheter-associated urinary tract infection (OR 0.97, 95% CI 0.52-1.80, I-2 = 0%, p = 0.55), and catheter-related blood stream infection (OR = 1.15, 95% CI 0.72-1.84, I-2 = 0%, p = 0.66), or ICU mortality (OR = 1.03, 95% CI 0.83-1.28, I-2 = 49%, p = 0.12). Forest plotting indicated that UVPs could reduce the lengths of ICU stays (SMD = - 0.97, 95% CI - 1.61 to 0.32, p = 0.003). Conclusion The current meta-analysis indicates that adopting a UVP may significantly reduce the incidence of delirium in ICU patients, without increasing the risks of ICU-acquired infection or mortality. Further large-scale, multicenter studies are needed to confirm these indications.