Scaphoid fractures are the most common carpal fractures. Diagnosing scaphoid fractures is challenging. Recently, cone-beam computed tomography (CBCT) has been shown to be a promising strategy for diagnosing scaphoid fractures. The diagnostic performance of CBCT remains inconclusive in the literature. Through a systematic review and meta-analysis, our study aims to determine the diagnostic performance of CBCT for diagnosing scaphoid fractures. Five databases were searched up to March 25, 2020. We included prospective and retrospective studies describing the diagnostic accuracy of CBCT for scaphoid fractures in adult patients. QUADAS-2 tool was used to assess the quality of the included studies. Four studies (n = 350) were included in the meta-analysis. Three of the four studies had high bias risk. The result showed that CBCT had a pooled sensitivity of 0.88 and a pooled specificity of 0.99 for scaphoid fracture diagnosis. The heterogeneities of sensitivity and specificity were substantial. The area under the summary receiver operating characteristic curve was 0.98. No significant publication bias was observed. The result suggested that the diagnostic performance of CBCT for scaphoid fracture was excellent. The certainty of current evidence is low. Further well-designed studies with large sample sizes are warranted to confirm this finding.