Fifteen young and 15 older healthy adults walked and crossed obstacles with heights of 10%, 20% and 30% of their leg lengths while their kinematic data were measured with a three-dimensional (3D) motion analysis system. End-point variables together with 3D joint kinematics of both the leading and trailing limbs were obtained. The results showed that the older group adopted a swing hip flexion strategy to achieve a higher leading toe clearance than the young group. With increasing obstacle height, the older group increased linearly the leading toe clearance by changing fewer joint angular components than the young group, allowing the maintenance of the necessary stability of the body with minimum control effort. When the trailing limb was crossing, the older group showed no significant difference in the trailing toe clearance compared to the young group, although different joint kinematic patterns were evident. The older group seemed to use a more conservative strategy for obstacle-crossing. Failure to implement this strategy during obstacle negotiation may increase the risk of falls owing to an inability to recover from unexpected tripping or stumbling. The results of this study suggest that existing knowledge of the kinematic control of obstacle-crossing based on young subjects may serve as a baseline for further studies on older people for a better understanding of the mechanisms and for the prevention of falls in the elderly.