Background: When maxillary transversal expansion is needed, two protocols of treatment can be used: a maxillary orthodontic expansion followed by a classical bimaxillary osteotomy or a bimaxillary osteotomy with maxillary segmentation. The aim of this study was to assess the accuracy of segmented Le Fort I osteotomy using computer-aided orthognathic surgery and patient-specific titanium plates in patients who underwent a bimaxillary osteotomy for occlusal trouble with maxillary transversal insufficiencies. Methods: A virtual simulation of a Le Fort I osteotomy with maxillary segmentation, a sagittal split ramus osteotomy, and genioplasty (if needed) was conducted on a preoperative three-dimensional (3D) model of each patient’s skull using ProPlan CMF 3.0 software (Materialise, Leuven, Belgium). Computer-assisted osteotomy saw-and-drill guides and patient-specific implants (PSIs, titanium plates) were produced and used during the surgery. We chose to focus on the maxillary repositioning accuracy by comparing the preoperative virtual surgical planning and the postoperative 3D outcome skulls using surface superimpositions and 13 standard dental and bone landmarks. Errors between these preoperative and postoperative landmarks were calculated and compared to discover if segmental maxillary repositioning using PSIs was accurate enough to be safely used to treat transversal insufficiencies. Results: A total of 22 consecutive patients—15 females and 7 males, with a mean age of 27.4 years—who underwent bimaxillary computer-assisted orthognathic surgery with maxillary segmentation were enrolled in the study. All patients presented with occlusion trouble, 13 with Class III malocclusions (59%) and 9 (41%) with Class II malocclusions. A quantitative analysis revealed that, overall, the mean absolute discrepancies for the x-axis (transversal dimension), y-axis (anterior–posterior dimensions), and z-axis (vertical dimension) were 0.59 mm, 0.74 mm, and 0.56 mm, respectively. The total error rate of maxillary repositioning was 0.62 mm between the postoperative cone-beam computed tomography (CBCT) and the preoperatively planned 3D skull. According to the literature, precision in maxilla repositioning is defined by an error rate (clinically relevant) at each landmark of <2 mm and a total error of <2 mm for each patient. Conclusions: A high degree of accuracy between the virtual plan and the postoperative result was observed.