Dancing is a complex sensorimotor activity involving physical and mental elements which have positive effects on cognitive functions and motor control. The present randomized controlled trial aims to analyze the effects of a dancing program on the performance on a motor-cognitive dual task. Data of 35 older adults, who were assigned to a dancing group or a health-related exercise group, are presented in the study. In pretest and posttest, we assessed cognitive performance and variability of minimum foot clearance, stride time, and stride length while walking. Regarding the cognitive performance and the stride-to-stride variability of minimum foot clearance, interaction effects have been found, indicating that dancing lowers gait variability to a higher extent than conventional health-related exercise. The data show that dancing improves minimum foot clearance variability and cognitive performance in a dual-task situation. Multi-task exercises (like dancing) might be a powerful tool to improve motor-cognitive dual-task performance.