The ability to generate peak power is central for performance in many sports. Currently two distinct resistance training methods are used to develop peak power, the heavy weight/slow velocity and light weight/fast velocity regimes. When using the light weight/fast velocity power training method it was proposed that peak power would be greater in a shoulder throw exercise compared with a normal shoulder press. Nine males performed three lifts in the shoulder press and shoulder throw at 30% and 40% of their one repetition maximum (1RM). These lifts were performed identically, except for the release of the bar in the throw condition. A potentiometer attached to the bar measured displacement and duration of the lifts. The time of bar release in the shoulder throw was determined with a pressure switch. ANOVA was used to examine statistically significant differences where the level of acceptance was set at p <0.05. Peak power was found to be significantly greater in the shoulder throw at 30% of 1 RM condition [F, (1, 23) =2.325 p <0.05) and at 40% of 1 RM [F, (1, 23) =2.905 p <0.05) compared to values recorded for the respective shoulder presses. Peak power was also greater in the 30% of 1 RM shoulder throw (510 +/- 103W) than in the 40% of 1 RM shoulder press (471 +/- 96W). Peak power was produced significantly later in the shoulder throw versus the shoulder press. This differing power reflected a greater bar velocity of the shoulder throw at both assigned weights compared with the shoulder press.