This study investigated the temporal organization during a mental rehearsal of 500 m speed skate sprint and a simple counting task without mental rehearsal. The effect of the difference between internal and external imagery (crassified by Mahoney and Avener, 1977) to time estimation in the mental rehenrsal was also investigated. Subjects were 8 elite speed skate athletes, ranging in age from 18 to 23 years, who had been fully trained by a image training program which included the same mental rehearsal. They were all highly motivated before each experiment. The mental rehearsal was the same to that in the previous study (Oishi et al. 1991).The mental rehearsal was started by a gun signal recorded by a tape recorder, after a rest period (5 min), relaxation (4 min), and a control of emotion (2.5 min). Subjects were instructed to start a watch at the gun signal and to stop the watch at their finish. Duration of the mental rehearsal was estimated after the each rehearsal. Another time estimation was done during the counting task, which was instructed to continue for the target time defineded as each best record of 500 m speed skate sprint without mental rehearsal. The estimate time during mental rehearsal was very close to the actual time. On the other hand, the other estimate time during counting task was deviated remarkably from the target time. In addition there were many different patterns of the imagery change from internal to external and vice versa. Though the internal imagery was dominant in our subjects, there was no obvious relation between the rate of internal imagery to external one and the accuracy of the time estimation during mental rehearsal. It was suggested that the mental movement was conducted with the similar velocity to that during the actual movement, and the central motor program was involved with the mental rehearsal.