This study investigated the effect of low- and high-velocity resistance training on isokinetic peak torque and anaerobic power output. Eighteen male varsity oarsmen were blocked on peak knee extension torque at 3.14 rad s-1 and assigned to a high-velocity resistance training group (HVR), a low-velocity resistance training group (LVR) or a control group. Subjects trained four times a week for 5 weeks. Each training session included three circuits of 12 stations using variable-resistance hydraulic equipment. The HVR training significantly improved peak torque (P less than 0.05) in knee extension and flexion at 2.61, 3.14, 3.66 and 4.19 rad s-1. The LVR training produced significant improvements (P less than 0.05) in peak torque for knee extension and flexion at 0.52, 1.05, 1.57 and 2.61 rad s-1. High positive correlations were found between peak torque and anaerobic power outputs for all groups. However, no significant changes occurred in 15 s power output, average 90 s power output or peak blood lactate in either training group. These results indicate that velocity-specific strength training does not necessarily improve anaerobic power output in a different exercise mode despite the high positive correlation between isokinetic strength and anaerobic power output.