The aim of this study was to assess, via an EMG bio-feedback method, the ankle joint angle effect on the agonist and antagonist torques in plantar- (PF) and dorsi-flexion (DF). The isometric PF and DF maximal voluntary contractions (MVCs) torques were measured simultaneously with surface EMG activity of triceps surae (TS) and tibialis anterior (TA) muscles in 12 young adults (mean age 27) at five different ankle joint angles. Our results showed that: (i) The coactivation level does not properly reflect the mechanical effect of the antagonist muscle, (ii) TS antagonist torque significantly altered the DF MVC–angle relationship, whereas TA antagonist torque did not influence this MVC–angle relationship in PF. The alteration of the MVC with angular position was due, in part, to the coactivation phenomenon in DF, but not in PF. Thenceforth, when investigating the torque at the ankle joint, it is necessary to take into account both agonist and antagonist torque modifications with ankle joint angle.