In the context of an experimental project requiring surgical preparation of animals, training is not only required for technicians lacking a strong surgical background but also for more experienced staff. The objectives of experimental surgery trainings are not only achieving good survival and success rates but also producing homogeneous animal batches, allowing reliable and reproducible responses to an experimental stimuli. The training program should be adapted to the surgeons experience and skills, as well as to the actual difficulty of surgical procedures to be performed. It should include a theoretical part that allows the trainee to understand the context of the surgery, identify the key anatomical structures and consequences of the operation on the animal's physiology. Strain, sex and age range specificities should also be taken in account. To insure adherence to the 3R's, ex-vivo models should be used as much as possible. Inanimate training tools or non-survival surgeries should be used whenever possible, unless a key component of surgical success requires live subjects. The number of key gestures repetitions required to reach training objectives described in the literature varies from 4 to more than 50 for the most complex models. Initial goals and associated number of required repetitions could be predefined but should remain flexible and adapted to the actual surgeon learning curve. Once the initial training goals are met, a continuous improvement/retraining program should be implemented. It should include controls that allow concomitantly step-by-step improvement and technical drift prevention.