Purpose: The purpose of this study was to assess the effects of robo-horseback riding with changes of pelvic tilting and speeds on muscle activities of trunk and lower limb in healthy adults. Methods: Thirty healthy adults were recruited for this study. Subjects performed robo-horseback riding at three pelvic postures (anterior tilting pelvic posture: AP, neutral pelvic posture: NP, posterior pelvic posture: PP) and different speeds (moderate and fast speed). Muscle activities were recorded from the trunk muscles (rectus abdominis [RA], T2 paraspinalis [T2 PS], L2 paraspinalis [L2 PS]) and lower limb muscles (rectus femoris [RF], vastus medialis [VM], vastus lateralis [VL], biceps femoris [BF], gastrucnemius [GCM]). Results: Significantly higher RA activities were observed under the PP condition than under the AP and NP conditions at both speeds. Significantly higher L2 PS activities were observed under the AP condition than under the PP and NP conditions at both speeds. Significantly higher T2 PS activities were observed under the AP condition than under the PP and NP conditions at fast speed. However, the muscle activities of the VM and GCM did not show significant changes with changes in pelvic posture at both speeds. All muscle activities were significantly higher under the fast speed than under the moderate speed in three pelvic postures. Conclusion: There were significant differences in the changes of pelvic tilting on muscle activities of trunk and lower limb were observed in healthy adults. NP condition can be introduced the most efficienct posture for muscle activities of trunk and lower limb in robo-horseback riding.