The present study examined the hypothesis that the antagonistic ankle dorsiflexor coactivation level during maximum isometric voluntary plantarflexion (MVC) is a function of ankle angle. Six male subjects generated plantarflexion and dorsiflexion MVC trials at ankle angles of -15 deg (dorsiflexed direction), 0 deg (neutral position), +15 deg (plantarflexed direction) and +30 deg having the knee flexed at an angle of 90 deg. In all contractions surface EMG measurements were taken from tibialis anterior and soleus which were considered representative muscles of all dorsiflexors and plantarflexors, respectively. Antagonistic dorsiflexor coactivation was expressed as normalized EMG and moment. Calculations of the antagonistic dorsiflexor moment were based on the tibialis anterior EMG-dorsiflexor moment relationship from contractions at 50, 40, 30, 20 and 10 % of the dorsiflexion MVC moment. In both legs dorsiflexor coactivation level followed an open U-shaped pattern as a function of ankle angle. Differences of 9 and 14 % (P < 0·05) were found in the measured net plantarflexion MVC moment between legs at ankle angles of -15 and +30 deg, respectively. No difference (P > 0·05) was found in the calf circumference between legs. Differences were found in the antagonistic dorsiflexor coactivation between legs at ankle angles of -15 and +30 deg. In the weaker leg the antagonistic EMG measurements were higher by 100 and 45 % (P < 0·01) and the estimated antagonistic moments were higher by 70 and 43 % (P < 0·01) compared with the weaker leg at -15 and +30 deg, respectively. This finding was associated with a decreased range of motion (ROM) in the weaker leg (14 %, P < 0·01), such that no difference (P > 0·05) was found in dorsiflexor antagonistic coactivation between legs at end-range ankle angles. The findings of the study (i) have to be taken into consideration when estimating musculoskeletal loads in the lower extremity, (ii) imply that stretching training can result in a stronger plantarflexion at end-range ankle angles through inhibition of the dorsiflexors, and (iii) imply a neural drive inadequacy during a plantarflexion MVC at end-range angles.