Malignant gastric outlet obstruction (GOO) can be palliated by endoscopic placement of self-expandable metal stents (SEMSs). Studies comparing uncovered (USEMSs) and covered (CSEMSs) SEMSs have yielded inconclusive results. This meta-analysis compared USEMSs with CSEMSs in patients with GOO caused by intrinsic and extrinsic tumors. Potentially relevant articles were identified by searching PubMed, the Cochrane Library, and Medline. The primary outcome was stent dysfunction. Relationships between characteristics of tumors causing GOO and the stent dysfunction rate following USEMSs and CSEMS placement were assessed. Overall, six randomized controlled trials and 12 observational studies, including 2431 patients, were identified. Rate of stent dysfunction did not differ significantly between USEMSs and CSEMSs (odds ratio [OR], 0.95; 95% confidence interval [CI], 0.73-1.25, I2 = 44%), but there was heterogeneity. Meta-regression analysis showed that heterogeneity of stent dysfunction was caused by a difference in the ratio of intrinsic-to-extrinsic tumors causing GOO between studies (coefficient, 0.944; 95% CI, 0.30-1.58). Studies were divided into intrinsic and extrinsic tumor groups. Subgroup analysis showed that the stent dysfunction rate did not differ between USEMSs and CSEMSs in the intrinsic tumor group (OR, 1.10; 95% CI, 0.85-1.41; I2 = 32%). In the extrinsic tumor group, USEMS was associated with a lower rate of stent dysfunction than CSEMS (OR, 0.64; 95% CI, 0.47-0.87; I2 = 25%). USEMS and CSEMS placement generally showed comparable outcomes among patients with GOO. USEMS was more effective than CSEMS for patients with GOO caused by extrinsic tumors.