Prismatic adaptation (PA) with wedge prisms is a non-invasive technique used in the rehabilitation of patients suffering from spatial neglect. Unfortunately, as for many behavioural intervention techniques, it is nearly impossible to achieve adequate blinding using wedge prisms, and the potential benefit of PA in the rehabilitation of neglect remains controversial. In order to study an alternative to wedge prism, we examine whether virtual PA at different degrees of deviation may alleviate signs of neglect in a double-blind design. Fifteen neglect patients participated in three adaptation sessions, which differed by the degree of deviation (0°, 15°, or 30°). Performance in line bisection and item cancellation tasks was measured in virtual reality immediately before and after adaptation. Session allocation was concealed from patients and the examiner. Despite the presence of robust, dose-dependent effects of virtual PA on Open-Loop Pointing (OLP), no transfer to line bisection and item cancellation tests were observed. None of the patients were aware of differences between sessions. Virtual PA did not result in visuo-motor transfer effects despite inducing significant adaptation effects in OLP. Together with recent negative findings of randomized-controlled trials, these findings cast doubt on the general efficacy of PA as a rehabilitation method of spatial neglect.