Patients with homonymous hemianopia (HH) are often crucially disabled during self-guided visual exploration of their natural environment. It is a matter of debate as to whether the abnormal visual search behaviour is related to a pure sensory deficit, a deficient spatial organization or individual compensation strategies. We tested the hypothesis that visual search behaviour in HH patients is purely determined by the sensory (visual) deficit by comparing them to healthy subjects with a technically simulated, virtual HH (VHH). Eye movements were recorded during an explorative visual search task in three groups: (i) nine patients with homonymous hemianopia due to unilateral posterior cerebral artery infarctions in their acute stage, (ii) nine age-matched healthy VHH subjects, and (iii) nine healthy control subjects. Subjects were asked to search for stationary visual targets on a monitor, which differed in colour and form from similar items serving as distractors. Identified targets had to be marked by a mouse click. The hemianopic field in VHH subjects was elicited by gaze-contingent blanking of one half of the visual hemifield, i.e. always with respect to the actual gaze position. Scanpath strategies did not differ between HH and VHH subjects. However, patients showed significantly longer search durations than VHH subjects, which was tightly related to their higher number of fixations and particularly re-fixations, the repeated scanning of previously fixated items. There was no increase in the number of re-clicks. Hemifield analyses within both groups, HH and VHH, did not show any differences concerning number of fixations and re-fixations or amplitude of saccades. Even patients with small lesions restricted to the visual cortex showed a prolonged search duration and higher re-fixation rate when compared to healthy VHH subjects. Patients' working memory and top-down scanning strategies seemed intact. Due to their acute stage, however, they did not develop adaptive eye movement strategies (no hemifield differences) as described in chronic HH patients after months. In conclusion, acute ischemic damage to the occipital lobe impairs visual search in HH patients to an extent, which the pure visual deficit does not fully account for. As scanpath stretegies and re-click rates were within normal limits and there was no visual neglect behavior, we cannot attribute these deficits to direct damage of higher cortical areas. Rather we assume that visual search abnormalities in HH are more likely due to impairment of connecting fibres to cortical (parietal) and subcortical (thalamical) brain regions.