The aim of this study was to identify the immediate effect of self-modelling with different focus of attention strategies (i.e., internal vs. external) on the teaching/learning of gymnastics motor-skills. Fifty-nine non-gymnast students participated in this study and were divided into three groups (i.e., an external focus group (EF), an internal focus group (IF), and a control group (CG)). Each participant’s performance of the back dismount in the parallel bars was assessed before the experiment (i.e., base-score), and each participant was asked to provide a self-evaluation of their performance and their efficiency percentage. Afterwards, participants received a specific learning session (i.e., self-modelling with external focus, self-modelling with internal focus, or traditional learning with verbal instruction) and performed the back dismount in the parallel bars again immediately after (i.e., final score). Four international judges evaluated performance of our participants. The results showed that the EF and IF outperformed the CG in the final score. Importantly, a significant difference between the base and the final score was observed in the EF and IF, but not in the CG. In addition, the EF showed the highest percentage of improvement (Δ-score) and self-estimation scores compared to the two other groups. In conclusion, this study supports the adoption of external focus of attention for teaching/learning gymnastics motor-skills.