The trainability of anticipatory skills for tennis was assessed. Subjects (N = 34) from a beginning/intermediate tennis class were randomly assigned to either a mental quickness or a physical quickness (control) training group. They were tested in three laboratory tennis simulation tasks and three on-court tasks (serves, ground strokes, and volleys) 1 week before and after the 3-week quickness training program. Quickness Training × Gender × Test Session (2 × 2 × 2) ANOVAs with repeated measures on the third factor were conducted. For the laboratory tasks, the mental quickness group made faster decisions in reaction to serves, exhibited faster anticipation times, and showed improved accuracy in predicting serve type and location. No improvements in accuracy were found for the physical quickness group. For filmed match-play situations, the mental quickness group improved reaction times with training and committed fewer response errors. Implications for the design of instructional methods used in dynamic and fast-paced sports are discussed.