Chronic ankle instability (CAI) has not previously been investigated dynamically using 3-D motion analysis during walking. Identification of altered kinematic and kinetic patterns in CAI subjects may increase our knowledge of the causes of CAI and improve our ability to prevent and successfully rehabilitate such subjects. We hypothesised that CAI subjects would exhibit a different kinematic and kinetic pattern of movement during normal walking when compared with a control group. Twenty five CAI subjects aged 25.5 ± 7.3 years (mean ± S.D.) and 25 age, activity, gender, and gait velocity matched controls, aged 23.61 ± 5.1 years volunteered to participate. Ethical approval was obtained locally and all subjects gave written consent. The CODA mpx30 3-D motion analysis system combined with a Bertec force plate were used to track joint motion and forces created during walking at natural self-selected velocity. Pelvic, thigh, shank, and foot 3-D segment joint angular rotation (subtracting joint angles when the subject was in their sub-talar neutral standing position), moments and powers during the period 100 ms pre-initial contact to 200 ms post-initial contact were identified for 10 trials for each subject. Average values for the above parameters were calculated for each subject and group mean profiles were derived. Independent two-sided t tests were used to analyse the data. P<0.05 was considered statistically significant. The CAI subjects during 100 ms pre-initial contact, and 200 ms post-initial contact were significantly (P<0.01) more inverted by approximately 6-7 deg in the frontal plane compared with the controls. From initial contact to 200 ms post-initial contact the joint moments show a significant difference (P<0.01) in the CAI subjects. The CAI subjects are controlled by an evertor moment compared with an invertor moment in the controls. The joint powers controlling CAI subjects from 10 ms to 40 ms, and 60 ms to 135 ms post-initial contact show a significant difference (P<0.03) to the controls. The CAI subjects are controlled by a concentric power generation compared with an eccentric power generation in the controls. These findings may indicate why subjects with CAI are vulnerable to repeated sprains. Landing on a more inverted foot during early stance may reduce the shock absorption mechanism, making the ankle more vulnerable to turning over on the lateral border during early stance. The findings suggest that the motor control of the ankle changes to prevent turning over on the lateral border of the foot in subjects with CAI during this period.