In the present study a method was developed to determine knee joint moment and power generated at low to very high velocities of knee extension. A group of 21 male subjects performed maximal knee extensions at four levels of external loading provided by a flywheel system. Knee extension was performed with no restrictions on joint angular velocity and acceleration. An interpolation procedure was employed to obtain moment and power at standard velocities. During each single knee extension peak velocity, peak moment, peak power and moment and power at 50 degrees knee flexion were determined (0 degree = full extension). While maximal recorded angular velocity averaged 693 degrees.s-1 (range 479-1030), maximal recorded moment and power were 226 N.m (157-278) and 1140 W (573-1688), respectively, which were generated at velocities of 55 degrees.s-1 (12-148) and 523 degrees.s-1 (327-846). Isokinetic moment and power were obtained for comparison. The velocity range applied was larger using the flywheel method (21-1030 degrees.s-1 individual range) than that achieved isokinetically (30-240 degrees.s-1). Significant differences were observed between flywheel and isokinetic measurements. These discrepancies may be ascribed to differences in the time course of muscle length change and contractile force generation imposed by the two measuring methods. However, by the nonisokinetic measuring method presently employed, knee extension moment and power could be determined from low to very high angular extension velocity, at loading and contraction conditions comparable to those encountered during sport and exercise.