Aims: To determine the efficacy of occipital nerve stimulation (ONS) in reducing the intensity, duration, and frequency of medically intractable primary headaches. Methods: A systematic review was carried out by searching three electronic databases: the Cochrane Library, MEDLINE via PubMed, and Web of Science. Randomized controlled trials (RCTs) and case series were eligible for inclusion. RCTs were assessed for quality of evidence by using the Cochrane Risk of Bias and the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) tools. Descriptive statistics of reported outcomes in eligible studies are presented in tabular form. Meta-analyses of RCTs comparing ONS therapy to sham therapy in chronic migraine patients were conducted for the outcomes responder rate, headache frequency, and headache intensity. Results: Four RCTs, 1 follow-up study, and 19 case series met the inclusion criteria. The quality of the evidence was low, with all four RCTs assessed as having a high risk of bias and small sample size. Meta-analyses of three RCTs showed patients receiving ONS therapy had a significant reduction of 3 headache days per month (difference in means = -3.061; 95% confidence interval [CI] = -5.162 to -0.961; P = .004) and a significant reduction in Migraine Disability Assessment score (standardized difference in means [SDM] = -0.634; 95% CI = -0.933 to -0.335; P .001) compared to sham (subthreshold) therapy. There were no statistically significant differences in reduction in pain intensity (SDM = -1.220; 95% CI = -2.489 to -0.049; P = .060) or in the number of responders (risk ratio [RR] = 1.581; 95% CI = 0.749 to 3.355; P = .229). Conclusion: ONS may be effective when compared to sham therapy, but the small number of RCTs and the heterogeneity of outcomes suggest further research in this field is needed. Keywords: chronic migraine, cluster headache, meta-analyses, occipital nerve stimulation, primary headache disorders