Objective: 3D-printed models hold great potential for temporal bone surgical training as a supplement to cadaveric dissection. Nevertheless, critical knowledge on manufacturing remains scattered, and little is known about whether use of these models improves surgical performance. This systematic review aims to explore (1) methods used for manufacturing and (2) how educational evidence supports using 3D-printed temporal bone models. Data Sources: PubMed, Embase, the Cochrane Library, and Web of Science. Review Methods: Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, relevant studies were identified and data on manufacturing and validation and/or training extracted by 2 reviewers. Quality assessment was performed using the Medical Education Research Study Quality Instrument tool; educational outcomes were determined according to Kirkpatrick’s model. Results: The search yielded 595 studies; 36 studies were found eligible and included for analysis. The described 3D-printed models were based on computed tomography scans from patients or cadavers. Processing included manual segmentation of key structures such as the facial nerve; postprocessing, for example, consisted of removal of print material inside the model. Overall, educational quality was low, and most studies evaluated their models using only expert and/or trainee opinion (ie, Kirkpatrick level 1). Most studies reported positive attitudes toward the models and their potential for training. Conclusion: Manufacturing and use of 3D-printed temporal bones for surgical training are widely reported in the literature. However, evidence to support their use and knowledge about both manufacturing and the effects on subsequent surgical performance are currently lacking. Therefore, stronger educational evidence and manufacturing knowhow are needed for widespread implementation of 3D-printed temporal bones in surgical curricula.