Objective: To evaluate the learning curve and safety of total thoracoscopic mitral valve repair (MVP). Background: Total thoracoscopic MVP is characterized by minimal trauma, minimal bleeding, and short postoperative recovery time. The learning curve of any new procedure needs to be evaluated for learning and replication. However, minimally invasive mitral valve technique is a wide-ranging concept, no further analysis of the outcomes and learning curve of total thoracoscopic mitral valve repair has been performed. Methods: One hundred and fifty consecutive patients who underwent minimally invasive MVP alone without concurrent surgery were evaluated. Using Cardiopulmonary bypass (CPB) time and Aortic clamping (AC) time as evaluation variables, we visualized the learning curve for total thoracoscopic MVP using Cumulative sum analysis. We also analyzed important postoperative variables such as postoperative drainage, duration of mechanical ventilation, ICU stay and postoperative hospital stay. Results: The slope of the fitted curve was negative after 75 procedures, and the learning curve could be crossed after the completion of the 75th procedure when AC and CPB time were used as evaluation variables. And as the number of surgical cases increased, CPB, AC, postoperative drainage, duration of mechanical ventilation, ICU stay and postoperative hospital stay all showed different degrees of decrease. The incidence of postoperative adverse events is similar to conventional mitral valve repair. Conclusions: Compared to conventional MVP, total thoracoscopic MVP provides the same satisfactory surgical results and stabilization can be achieved gradually after completion of the 75th procedure.