The aim of the present work was to analyze the influence of self-controlled task difficulty on motor learning. Participants had to intercept three targets falling at different velocities by displacing a stylus above a digitizer. Task difficulty corresponded to racquet width. Half the participants (self-control condition) could choose the racquet width at the beginning of each trial. Each was paired with a participant from the yoked group. The self-control condition resulted in better performances and accuracy during immediate and delayed retention tests. These results confirm the advantage of a self-control condition on motor learning. They are discussed with reference to the challenge point hypothesis (Guadagnoli & Lee, 2004).