This paper presents a gait analysis in the rehabilitation walking of a hemiplegic patient with stroke. Temporal and spatial parameters during gait are calculated from footprint images recorded on a paper of 10m length, and acceleration data. Gait characteristics of two subjects are discussed with a focus on the progress of rehabilitation. In order to provide the degree of recovery, it is shown that it is effective to investigate the changes in the movement of body trunk and the length of shuffling gait. Moreover, this study examines gait improvement by wearing plastic ankle foot orthosis and considers the validity of the support function.