A computer model of the musculoskelotal system that provides accurate prediction of muscle force and body movement trom the stimulation input is desired for the effective control system design in FES. This paper aims to investigate the fundamental properties of the gradual muscle force potentiation that was not included in the previous muscle models, for future development of a model that provides vetter prediction of FES-induced muscle force and body movement. Specifically, hou the muscle length was investigated. The experimental results showed that both the force increment ratio and the time-to-peak during electrical stimulation decreased with stimulatino frequency. When the muscle potentiation state was saturated by preceding stimulation. the force did not increase any more during additive stimulation. Muscle length significantly affected the force potentiation in such a way that the force increment ratio decreased with muscle length. A new model of the muscle potentiation based on these results is desired in the future.