This study aimed to investigate the role of rest period duration (RP) on the time course of the acceleration portion (AP) and mean velocity of the concentric phase across two bench press sets to failure with a submaximal load (60% of the 1RM) using different RP. Ten elite junior kayakers performed, on four different days, two consecutive bench press sets to failure, allowing randomly 1-, 2-, 3- and 4-min RP between sets. AP reached a maximal value of 66% of the concentric movement time. This maximal AP was observed in repetition number 2 or 3, and then AP declined during the set, with a significant decrease when the number of repetitions was over 80% of the total number of repetitions performed. AP and lifting velocity patterns of the concentric phase were not altered during a second set to failure, regardless of RP. However, when velocity was expressed in absolute terms, 1-min RP was insufficient to maintain the average lifting velocity during the second set, compared to the first one. These results may be of use in selecting number of repetitions and resting duration in order to ensure optimal maintenance of the accelerative portion of concentric movement time with different resting-period durations.