Background: The incidence of children developing recurrent sacrococcygeal teratoma (SCT) is 2–35%. Serum alpha-fetoprotein (AFP) is often used as a tumor marker for (malignant) recurrences of SCT and could potentially be used during routine follow-up after SCT resection. However, the diagnostic accuracy of serum AFP levels during follow-up has not been well established. Therefore, we aimed to systematically review the diagnostic accuracy of serum AFP levels in recurrent SCT. Methods: We queried Search Premier, COCHRANE Library, EMCARE, EMBASE, PubMed, ScienceDirect and Web of Science databases to identify studies regarding patients with SCT with follow-up using serum AFP levels postoperative. We estimated sensitivity and specificity of serum AFP levels. Results: Fifteen studies (613 patients, 121 recurrences) were included and these mainly described serum AFP levels in patients with recurrent SCT (n = 111); 83 (75%) patients with recurrent SCT had elevated serum AFP levels. A subgroup analysis of articles that measured serum AFP levels in all patients (n = 6, 136 patients, 14 recurrences) showed a sensitivity and specificity of 79% and 95%, respectively. The sensitivity of AFP levels to detect malignant recurrence was 96%. Conclusion: Diagnostic accuracy of serum AFP levels to detect recurrent SCT seems promising, though sensitivity could be overestimated since serum AFP levels are mainly described in patients with elevated AFP levels or at recurrent SCT. Furthermore, serum AFP levels could be helpful to detect malignant recurrences. Type of study: Systematic review of level 2–4 studies. Level of evidence: Level 2–4 (mostly level 2).