To perform a systemic review and meta-analysis of the diagnostic accuracy of PET (CT) and metaiodobenzylguanidine (MIBG) for diagnosing neuroblastoma (NB), electronic databases were searched as well as relevant references and conference proceedings. The diagnostic accuracy of MIBG and PET (CT) was calculated for NB, primary NB, and relapse/metastasis of NB based on their sensitivity, specificity, and area under the summary receiver operating characteristic curve (AUSROC) in terms of per-lesion and per-patient data. A total of 40 eligible studies comprising 1134 patients with 939 NB lesions were considered for the meta-analysis. For the staging of NB, the per-lesion AUSROC value of MIBG was lower than that of PET (CT) [0.8064 +/- 0.0414 vs. 0.9366 +/- 0.0166 (P < 0.05)]. The per-patient AUSROC value of MIBG and PET (CT) for the diagnosis of NB was 0.8771 +/- 0.0230 and 0.6851 +/- 0.2111, respectively. The summary sensitivity for MIBG and PET (CT) was 0.79 and 0.89, respectively. The summary specificity for MIBG and PET (CT) was 0.84 and 0.71, respectively. PET (CT) showed higher per-lesion accuracy than MIBG and might be the preferred modality for the staging of NB. On the other hand, MIBG has a comparable diagnosing performance with PET (CT) in per-patient analysis but shows a better specificity.