Background Surgeons are pursuing accurate head and neck reconstruction to enhance aesthetic and functional outcomes after oncologic resection. This study aimed to investigate whether accuracy of head and neck reconstruction is improved with the use of three-dimensionally (3D)-printed patient-specific surgical plates compared with conventional plates. Methods In this comparative study, patients were prospectively recruited into the study group (3DJP16) with 3D-printed patient-specific surgical plates. The patients in control group with conventional surgical plates were from a historic cohort in the same unit. The primary end point of the study was the accuracy of head and neck reconstruction. The secondary end points were accuracy of osteotomy, intraoperative blood loss, total operative time, and length of hospital stay. Results The study recruited of 33 patients, including 17 in the study group and 16 in the control group. The patients’ baseline characteristics were similar between the two groups. The absolute distance deviation of the maxilla or mandible was 1.5 ± 0.5 mm in the study group and 2.1 ± 0.7 mm in the control group [mean difference, − 0.7 mm; 95% confidence interval (CI) − 1.1 to − 0.3; p = 0.003], showing superior accuracy of reconstruction for the patients with 3D-printed patient-specific surgical plates. Improved accuracy of reconstruction also was detected in terms of bilateral mandibular angles and bone grafts. Concerning the secondary end points, the accuracy of the osteotomy was similar in the two groups. No difference was found regarding intraoperative blood loss, total operative time, or length of hospital stay. Conclusions This is the first study to prove that compared with conventional plates, 3D-printed patient-specific surgical plates improve the accuracy of oncologic head and neck reconstruction.