In this study we examined (a) whether verbal self-instructions can enhance task-switching performance in younger and older adults, and (b) whether verbal self-instruction benefits on task switching are smaller when memory demands on keeping track of the task sequence are reduced by spatial task cueing. Task-switching ability was measured as the difference in performance between single-task and mixed-task blocks (termed mixing costs), in which participants switched between two tasks A and B. One group of participants performed the switching tasks with spatial task cues, indicating which of the two tasks has to be performed, thereby reducing demands on the endogenous control of serial task order (low task-sequencing load). The other group switched between tasks without external task cues (high task-sequencing load). To investigate the influence of verbal self-instructions on task switching, participants either named aloud the next task during task preparation (task-naming condition) or they did not verbalize (control condition). Results indicated that age differences in verbalization benefits on mixing costs depend on early learning whereby benefits were generally larger when subjects had some prior practice in task switching alone, and that verbalization benefits did not differ between the two task-sequencing load conditions. These findings suggest that task naming is a suitable cognitive intervention for enhancing the control of task switching in younger and older adults, even if memory load is reduced, and that for the efficient application of this strategy it first has to be coordinated with task switching, which is easier when task switching is already practiced.