The investigation examined isokinetic (IK) and nonisokinetic (NIK) strength training programs for the inversion (INV) and eversion (EV) muscles on pronation during running. Seventy-seven volunteers were videotaped running on a treadmill at 3.8 m.s-1 and total pronation (delta beta PRO) was computed. Eighteen heel-strike runners with the largest values of delta beta PRO (X = 16.7 degrees) were selected as subjects. During the pre- and posttests, isokinetic muscle strength at 20 and 180 degrees.s-1 was determined for the concentric (CON) and eccentric (ECC) actions of the INV and EV muscle groups. The subjects also were videotaped running on a treadmill (3.8 m.s-1). The IK training group performed three sets of eight CON and ECC repetitions at 20, 90, and 180 degrees.s-1 for both muscle groups; and the NIK subjects did exercises commonly used in ankle rehabilitation. Each group trained three times weekly for 8 wk. The IK group showed significantly (P < or = 0.05) CON and ECC strength increases for all INV test conditions and three of the four EV conditions (20 degrees.s-1 CON and ECC, and 180 degrees.s-1 CON). They also demonstrated significant decreases in the rearfoot (2.2 degrees) and pronation/supination (2.9 degrees) angles at heel strike and in delta beta PRO (-2.2 degrees).l The NIK group exhibited no change in rearfoot motion and only increased INV strength at the 180 degrees.s-1 ECC test condition. The findings suggest that pronation can be decreased by an isokinetic strength training program for the INV and EV muscles.