We examined retrieval-induced forgetting of eye movements. The items were intentionally encoded letter-cued eye movements, indicated by a moving box over a 6x6 grid on a computer screen display. These directionally categorized movements were animated across the screen from left-to-right or vice versa. Participants had to track the movements with their eyes, while remembering both, corresponding movement and letter cue. The learning phase encompassed six letter-cued eye movements over this pattern from left to right and six letter-cued movements in the opposite direction. After sufficient initial learning of these twelve eye movement sequences belonging to their categorial cue, retrieval-practice started. Half of the items of one movement category (e.g. left-to-right) were retrieval-practiced several times. In a subsequent cued recall test phase in Experiment 1, participants had to explicitly remember all letter-cued movements in a given order. Results showed retrieval-induced forgetting in an accuracy measure of recall, that is, significantly lower recall of non-retrieval-practiced eye movements belonging to the retrieval-practiced category as compared to the non-practiced category. In Experiment 2, the final test phase encompassed an indirect test. Here, participants were not demanded to remember the letter-cued sequences, instead they had to execute these movements by following a new indicator on the screen, representing the eye movement in an indirect manner. Results here also showed retrieval-induced forgetting of eye movements, this time in significant longer execution time for the inhibited movements. In summary, we demonstrated retrieval-induced forgetting of eye movements in a direct (Experiment 1) and indirect (Experiment 2) testing procedure.