The preservation of perception-action coupling in the practice environment may increase skill transfer to competition and quality of learning, in part because relevant perceptual (e.g., visual) information is preserved. Although baseball batters rely on advanced gaze behaviours to achieve success in a challenging perceptual-cognitive-motor task, development of these skills may be constrained by traditional de-coupled practice drills. This study's purpose was to explore the influence of de-coupled batting drills on the gaze behaviour of 28 elite baseball players. Participants completed one of four warm–up tasks common in baseball practice (i.e., stationary tee, pitching machine batting practice, coach-thrown batting practice, and a control) varying in representativeness, prior to participating in a temporal occlusion simulation. No statistically significant evidence of an acute warm-up effect was noted. However, descriptively the more de-coupled tasks influenced participants' quickness and ability to locate the pitcher's release point and to less desirable observations of task-relevant areas. Main effects of occlusion time (F (2, 26) = 3.87, p = .03), and playing level (F (1, 27) = 16.81, p < .01), were also observed. Accordingly, players of advanced skill fixated more quickly on the release point, and participants increased their sum of task-relevant fixations across trials. Additionally, release point fixation before release of the ball was positively related to correct pitch type predictions (?2 (1) = 78.67, p < .01). The absence of a conclusive warm-up condition effect may be explained by the athletes' familiarity with unrepresentative tasks, and potential skill in recalibration.