The aim of the present study was to investigate the EMG-joint angle relationship during voluntary contraction with maximum effort and the differences in activity among three hamstring muscles during knee flexion. Ten healthy subjects performed maximum voluntary isometric and isokinetic knee flexion. The isometric tests were performed for 5 s at knee angles of 60 and 90 degrees. The isokinetic test, which consisted of knee flexion from 0 to 120 degrees in the prone position, was performed at an angular velocity of 30 degrees /s (0.523 rad/s). The knee flexion torque was measured using a KIN-COM isokinetic dynamometer. The individual EMG activity of the hamstrings, i.e. the semitendinosus, semimembranosus, long head of the biceps femoris and short head of the biceps femoris muscles, was detected using a bipolar fine wire electrode. With isometric testing, the knee flexion torque at 60 degrees knee flexion was greater than that at 90 degrees. The mean peak isokinetic torque occurred from 15 to 30 degrees knee flexion angle and then the torque decreased as the knee angle increased (p<0.01). The EMG activity of the hamstring muscles varied with the change in knee flexion angle except for the short head of the biceps femoris muscle under isometric condition. With isometric contraction, the integrated EMGs of the semitendinosus and semimembranosus muscles at a knee flexion angle of 60 degrees were significantly lower than that at 90 degrees. During maximum isokinetic contraction, the integrated EMGs of the semitendinosus, semimembranosus and short head of the biceps femoris muscles increased significantly as the knee angle increased from 0 to 105 degrees of knee flexion (p<0.05). On the other hand, the integrated EMG of the long head of the biceps femoris muscle at a knee angle of 60 degrees was significantly greater than that at 90 degrees knee flexion with isometric testing (p<0.01). During maximum isokinetic contraction, the integrated EMG was the greatest at a knee angle between 15 and 30 degrees, and then significantly decreased as the knee angle increased from 30 to 120 degrees (p<0.01). These results demonstrate that the EMG activity of hamstring muscles during maximum isometric and isokinetic knee flexion varies with change in muscle length or joint angle, and that the activity of the long head of the biceps femoris muscle differs considerably from the other three heads of hamstrings.