Technology that assists and extends various functions of human beings will soon be available not only to medical and welfare but also to healthy individuals. This study aimed to characterize surface electromyography (EMG) signals in response to walking assistive equipment. Ten healthy male students walked on an 8-m uphill road (5.8% incline) using an assist walker (RT.2, RT.WORKS Co., Ltd) under assist and non-assist conditions. The EMG signals were recorded from four muscles (the rectus femoris [RF], biceps femoris, tibialis anterior, and lateral gastrocnemius). During loading response and terminal stance, the percent maximum voluntary isometric contraction (%MVC) peak value for RF was achieved more quickly in the assist condition than in the non-assist condition. However, during loading response and mid-swing, the %MVC peak value of RF was significantly lower in the assist condition than in the non-assist condition. These results indicate that humans alter muscle exertion patterns in specific muscles to adapt to walking assistance; such a change in the muscle exertion pattern may be adapted for smoother walking.