The purpose of this study was to investigate the effect of gait speed and walkway slope on the body acceleration, for the future validation of using an accelerometer in the estimation of energy consumption. Ten young healthy subjects with accelerometers on the upper thigh and ankle walked on a treadmill at 9 conditions(three speeds ${ imes}$ three slopes) for 5 minutes. Acceleration signals of four directions, i.e. anterior-posterior(AP), medio-lateral(ML), superior-inferior(SI) and vector sum(VS) directions, of each sensor were measured, and root means squared(RMS) values of them were used as analysis variables. As statistical analysis, repeated measure two-way ANOVA was performed for RMS accelerations at each attachment sites, with slope and velocity as independent factors. At both the upper thigh and ankle, RMS acceleration of all directions were affected by gait velocities(p<.001) showing greater accelerations for higher velocities. Contrary to expectations, no slope effect existed in RMS accelerations at hip. Moreover, RMS acceleraion at ankle decreased with slope in SI and VS directions(p<.01). These results suggests that RMS acceleration cannot reflect the change in physical activity due to the change in walkway slope.