The purpose of this study was to investigate chaotic characteristics of major joint motions during treadmill walking. Gait experiments were carried out for 20 healthy young women. The subjects were asked to walk on a treadmill at their own natural speeds. The chaos analysis was used to quantify nonlinear motions of eleven major joints of each woman. The joints analyzed included the neck and the right and left shoulders, elbows, hips, knees and ankles. The recorded gait patterns were digitized and then coordinated by motion analysis software. Lyapunov exponent for every joint was calculated to evaluate joint characteristics from a state space created by time series and its embedding dimension. This study shows that differences in joint motion were statistically significant.