The effects of mental fatigue on planning and preparation for future actions were examined, using a task switching paradigm. Fatigue was induced by “time on task,” with subjects performing a switch task continuously for 2 hr. Subjects had to alternate between tasks on every second trial, so that a new task set was required on every second trial. Manipulations of response–stimulus intervals (RSIs) were used to examine whether subjects prepared themselves for the task change. Behavioral measurements, event-related potentials (ERPs), and mood questionnaires were used to assess the effects of mental fatigue. Reaction times (RTs) were faster on trials in which no change in task set was required in comparison with switch trials, requiring a new task set. Long RSIs were used efficiently to prepare for the processing of subsequent stimuli. With increasing mental fatigue, preparation processes seemed to become less adequate and the number of errors increased. A clear poststimulus parietal negativity was observed on repetition trials, which reduced with time on task. This attention-related component was less pronounced in switch trials; instead, ERPs elicited in switch trials showed a clear frontal negativity. This negativity was also diminished by time on task. ERP differences between repetition and switch trials became smaller with increasing time on task.