The aim of this study was to compare the effects of a weight training program for the leg extensors with isokinetic cycling training (80 rpm) on maximal power output and endurance performance. Both strength training interventions were incorporated twice a week in a similar endurance training program of 12 weeks. Eighteen trained male cyclists (VO(2peak) 60 +/- 1 ml kg(-1) min(-1)) were grouped into the weight training (WT n = 9) or the isokinetic training group (IT n = 9) matched for training background and sprint power (P (max)), assessed from five maximal sprints (5 s) on an isokinetic bicycle ergometer at cadences between 40 and 120 rpm. Crank torque was measured (1 kHz) to determine the torque distribution during pedaling. Endurance performance was evaluated by measuring power, heart rate and lactate during a graded exercise test to exhaustion and a 30-min performance test. All tests were performed on subjects' individual race bicycle. Knee extension torque was evaluated isometrically at 115 degrees knee angle and dynamically at 200 degrees s(-1) using an isokinetic dynamometer. P (max) at 40 rpm increased in both the groups (~15%; P < 0.05). At 120 rpm, no improvement of P (max) was found in the IT training group, which was possibly related to an observed change in crank torque at high cadences (P < 0.05). Both groups improved their power output in the 30-min performance test (P < 0.05). Isometric knee extension torque increased only in WT (P < 0.05). In conclusion, at low cadences, P (max) improved in both training groups. However, in the IT training group, a disturbed pedaling technique compromises an improvement of P (max) at high cadences.