Objective: To analyze predictors that influence the learning curve of minimally invasive mitral valve surgery (MIMVS). Methods: Patients who underwent MIMVS between March 2010 to March 2015 were retrospectively analyzed. Predictive factors that influence the learning curve were analyzed. Results: One hundred and five patients were included in the analysis. Cardiopulmonary bypass (CPB) time in minutes was 158.72 ± 40.98 and the aortic crossclamp (ACC) time in minutes was 114.48 ± 27.29. There were three operative mortalities, one stroke and five >2+ mitral regurgitation. ACC time in minutes was higher in the low logistic Euroscore II (LES) group (LES < 5% = 118.42 ± 27.94) versus (LES ≥ 5 = 88.66 ± 22.26), P < .05 while creatinine clearance in μmol/L was higher in the LES < 5% group (LES < 5% = 84.32 ± 33.7) versus (LES ≥ 5% = 41.66 ± 17.14), (P < .05). One patient from each group required chest tube insertion for pleural effusion P < .05. The cumulative sum analysis (CUSUM) for the first 25 patients had CPB and ACC times that reached the upper limits. Between 25 to 64 patients the curve remained stable while with the introduction of reoperations and complex surgical procedures the CUSUM reached the upper limits. Conclusions: The learning curve is affected by many factors but this should not desist surgeons from approaching this technique. The introduction of high-risk patients in clinical practice should be carefully measured based on surgeon experience.