Human bipedal locomotion is an automatic function that is important for daily living. Modeling the fundamental sensorimotor strategies associated with walking and turning could be important for diagnosing diseases that affect gait and devising prostheses to aid in deficient locomotion. It would also be important for design of more efficient robots that could maneuver over uneven terrain, which is quite challenging for robots with wheels or many-legged robots. We have developed a model of reflex head and eye movements generated by the vestibular system (system that senses head rotation and head tilts with regard to gravity), which we have postulated as an important component for maintining stable gait. We have tested our model of locomotion using a video based system from Northern Digital, Inc (Optotrak) that measures head and body movements in real time. Our results indicate that during straight walking, the trunk and head rotations are coordinated to maintain an approximately invariant point in space associated with the intersection of sequences of roll axes of the head coordinate frame ( lines emanating from the back of the head to the nose) relative to trajectory coordinates. During turning, there is an additional orientation mechanism in the central nervous system that tends to align the head yaw axis (an axis emanating from the top of the head) with the changing net gravito-inertial acceleration (GIA). As the head rotates on the trunk, the eyes also rotate in a compensatory manner and orient gaze to the net GIA. Together the head and eye movement systems exercise top-down control to steer trajectories, maintain gaze stability during walking and turning and allow the body to maneuver efficiently without falling.