The precise role of cognitive control in the processing of optic flow has been rarely investigated. Therefore, this study aimed to determine whether coping with unreliable visual inputs during walking requires cognitive resources. Twenty-four healthy young adults walked on an instrumented treadmill in a virtual environment under two optic flow conditions: normal (congruent) and perturbed (continuous mediolateral pseudo-random oscillations). Each condition was performed under single-task and dual-task conditions of increasing difficulty (1-, 2-, 3-back). Foot placement kinematics (200 Hz) and surface electromyography (EMG) of soleus and gluteus medius (1000 Hz) were recorded. Means, standard deviations (variability), statistical persistence and step-to-step error correction were computed from gait time series in lateral and anteroposterior directions. For EMG variables, duration and variability of muscle activation were calculated from the full width at half maximum (FWHM) and the variance ratio, respectively. N-back task performance was expressed by the d prime and response time. Cognitive performance decreased as cognitive load increased (p<.001) but remained preserved under dual-task walking. Kinematics variability and EMG variance ratio increased under optic flow perturbation (p<.001). However, dual-tasking reduced the impact of the optic flow disturbance on the kinematics variability. Persistence of step width and antipersistence of step velocity decreased as cognitive load increased. Lastly, FWHM of soleus muscle increased with dual-task (p=.01). The results indicated that cognitive dual-tasking decreased the optic flow effect, demonstrating that resolving conflicting visual cues was attentionally costly. Also, in dual-task conditions, individuals adopted a rigid and controlled walking pattern in order to succeed N-back task.