A cluster-randomized controlled trial was conducted to examine the effects of a 4-week program that integrated movements into cognitive tasks related to numerical skills. Participants (N = 120, M age = 4.70 years, SD = 0.49; 57 girls) were assigned to one of the following four conditions: performing integrated physical activity (task relevant), performing nonintegrated physical activity (task nonrelevant), observing integrated physical activity, or conventional sedentary teaching (without performing or observing physical activity). Results showed that children who performed task-relevant integrated physical activity performed better than children in all other conditions. In addition, children who performed physical activity, either integrated or nonintegrated, reported higher scores for enjoyment of the instructional method than the two sedentary learning conditions. Implications for educational theory and practice are discussed.