To determine the effect of self-talk on softball throwing performance. Additionally, two moderators, nature of self-talk and type of motor task, as well as a potential mediator of self-efficacy were examined.An experimental, within-subjects, and counterbalanced design.Forty-two senior high students (mean age = 17.48 ± 0.55) were instructed to use instructional, motivational, and unrelated self-talk with counterbalanced order prior to softball throwing for accuracy and distance tasks.Both instructional and motivational self-talk conditions had better performance than unrelated self-talk on softball throwing accuracy, whereas motivational self-talk had better performance than both instructional and unrelated self-talk in softball throwing for distance. Results for self-efficacy were similar, with self-efficacy for accuracy performance higher in both instructional and motivational self-talk conditions than with unrelated self-talk, while self-efficacy was highest in the motivational self-talk condition and lowest with unrelated self-talk. Significant correlations between self-efficacy and motor performance were also found with both tasks.These findings partially support the task-matching hypothesis, confirm the moderator role of type of self-talk and task type, suggest that self-efficacy has a mediator role, and provide direction for self-talk effectiveness.