Objective: To explore the application accuracy of virtual preoperative plan after the condylectomy via intraoral approach under computer assisted surgical navigation, and to analyze the location and cause of the surgical deviation to provide reference for the surgical procedure improvement in the future. Methods: In the study, 23 cases with condylar hypertrophy (11 with condylar osteochondroma and 12 with condylar benign hypertrophy) in Department of Oral and Maxilloficial Surgery, Peking University School and Hospital of Atomatology from December 2012 to December 2016 were treated by condylectomy via intraoral approach under computer assisted surgical navigation. The patient’s spiral CT data were imported into ProPlan software before operation, and the affected mandibular ramus was reconstructed three-dimensionally. The condylar osteotomy line was designed according to the lesion range, and the preoperative design model was generated and introduced into the BrainLab navigation system. Under the guidance of computer navigation, the intraoral approach was used to complete the condylar resection according to the preoperative design of the osteotomy line. Cranial spiral CT of the craniofacial region was taken within one week after operation. Three-dimensional reconstruction of the mandibular ramus at the condylectomy side was performed, and the condylar section was divided into six segments (anterolateral, anterior, anteromedial, posteromedial, posterior, and posterolateral) and the corresponding regional measurement points P1 to P6 were defined. Then the preoperative virtual model and the postoperative actual model were matched by Geomagic studio 12.0 to compare the differences and to analyze the accuracy of the operation. Results: All the patients had successfully accomplished the operation and obtained satisfactory results. Postoperative CT showed that the condyle lesion was completely resected, and the condylar osteotomy line was basically consistent with the surgical design. No tumor recurrence or temporomandibular joint ankylosis during the follow-up period. The postoperative accuracy analysis of the condylar resection showed that the confidence intervals measured by the six groups of P1 to P6 were(-2.26 mm, -1.89 mm), (-2.30 mm, -1.45 mm),(-3.37 mm, -2.91 mm),(-2.83 mm, -1.75 mm),(-1.13 mm, 0.99 mm), and(-1.17 mm, 0.17 mm), where P3 group was different from the other 5 groups. There was no significant difference between the P5 and P6 groups and the difference between the other four groups was statistically significant. Conclusion: Under the guidance of computer navigation, the intraoral approach can be performed more accurately. The surgical deviation of each part of the osteo-tomy surface is mainly due to excessive resection. The anterior medial area of the anterior medial condyle represents the most excessive resection. The posterior and posterior lateral measurement points represent the posterior condylar area. The average deviation is not large, but the fluctuation of the deviation value is larger than that of the other four groups. The accuracy of computer-assisted subtotal resection has yet to be improved.