The accuracy of orthognathic surgery has improved with three-dimensional virtual planning. The translation of the planning to the surgical result is reported to vary by >2mm. The aim of this randomized controlled multi-centre trial was to determine whether the use of splintless patient-specific osteosynthesis can improve the accuracy of maxillary translation. Patients requiring a Le Fort I osteotomy were included in the trial. The intervention group was treated using patient-specific osteosynthesis and the control group with conventional osteosynthesis and splint-based positioning. Fifty-eight patients completed the study protocol, 27 in the patient-specific osteosynthesis group and 31 in the control group. The per protocol median anteroposterior deviation was found to be 1.05mm (interquartile range (IQR) 0.45–2.72mm) in the patient-specific osteosynthesis group and 1.74mm (IQR 1.02–3.02mm) in the control group. The cranial–caudal deviation was 0.87mm (IQR 0.49–1.44mm) and 0.98mm (IQR 0.28–2.10mm), respectively, whereas the left–right translation deviation was 0.46mm (IQR 0.19–0.96mm) in the patient-specific osteosynthesis group and 1.07mm (IQR 0.62–1.55mm) in the control group. The splintless patient-specific osteosynthesis method improves the accuracy of maxillary translations in orthognathic surgery and is clinically relevant for planned anteroposterior translations of more than 3.70mm.