OBJECTIVES Minimally invasive mitral valve surgery (MIMVS) through an endoscopic right minithoracotomy is a well-established yet complex procedure that has a challenging learning curve. We have developed a simulator for MIMVS and evaluated its short- and long-term training effects. METHODS Trainees without simulator experience or training in MIMVS were divided into 2 groups (10 students and 10 residents) and participated in a 5-day training course after initial instruction. Each trainee performed a ring annuloplasty. Scores were given by a supervisor who assessed 5 skills. The duration of each procedure was also measured. To evaluate the long-term effect of the training course, trainees performed the same procedure 4 weeks after the last session. RESULTS Trainees in the resident group were significantly older compared to those in the student group and had a mean surgical experience of 4.4 ± 0.78 years standard error of the mean. All other demographic data were similar. Significant learning curves could be achieved in both groups over the course of 5 days with regard to total skill scores and total duration. However, when we compared the learning curves of both groups, no significant difference could be seen. Long-term performance in both groups was still significantly better compared to that in the first training session. CONCLUSIONS Training with our simulator provided a significant enhancement of a trainee’s performance. This learning effect was achieved in both groups and was still evident 4 weeks later. We strongly recommend our simulator for simulation-based surgical education of cardiac surgeons interested in MIMVS.