Objective To investigate the assembl ing and cl inical appl ication of the video output system util izing teaching sight glass of surgical microscope. Methods Between June 2009 and April 2010, 10 patients with craniocervical junction malformation were treated by the method of transoral-transpharyngeal approach with the microscope and videooutput system under the direct vision. There were 6 males and 4 females with an average age of 32 years (range, 13-52 years). Three cases had the history of injury and 7 cases had no history of definite injury. The disease duration was from 10 months to 12 years (median, 5 years). The main cl inical symptoms were brevicoll is or torticoll is; 2 patients had malformation appearance and 4 patients had occi put-cervical pain. The physical examination showed that all patients had the symptoms that upper cervical cord was damaged; the imaging examination showed that all patients had basilar invagination, atlantoaxial dislocation, and ossification. Before and after operations, the functions of nerve were evaluated by Japanese Orthopaedic Association (JOA) scoring, the improvement rate was calculated to evaluate the efficacy. Results By the video output system assembly, 15.1 mill ion pixels high-definition images could be collected and reached 1 920 × 1 080 pixels video camera, so assistants or medical students could watch the cl inical operation directly. All patients had no neural injury or cerebrospinal fluid leakage during operation. Basilar invagination and atlantoaxial dislocation were corrected. Infection at incision occurred in 1 patient; other incisions healed by first intention without early compl ication. All patients were followed up 6-16 months (mean, 13.5 months). The average JOA score was increased from 10.2 preoperatively to 15.5 at 6 months postoperatively with an improvement rate of 77.9%. At 12 months after operation, bony fusions were achieved. Conclusion The miscroscope and video output system can improve the effectiveness of the original surgical microscope. It makes visual fields much clearer and operations more accuratewith a few compl ications.