Mind-wandering during reading is often associated with worse comprehension performance. Research suggests that metacognitive competences (i.e., the knowledge about as well as monitoring and regulation of one's own cognitive processes) are positively related to reading performance and may play a role in the prevalence of mind-wandering. However, whether metacognitive training can reduce task-unrelated thoughts and their negative impact on comprehension performance has yet to be investigated. The main objective of the present study was to investigate whether metacognitive training is effective in reducing participants' task-unrelated thoughts during reading. Participants (N = 80 undergraduate students) were randomly assigned to the metacognitive training or control training condition. The effects of the interventions were examined by measuring self-reported mind-wandering, unusually fast or slow reading times as an index for inattentive reading, and comprehension on the textbase and situation model level before and after training. Results revealed less self-reported mindless reading and fewer instances of unusually fast or slow reading times after the metacognitive training but not after the control training. The metacognitive condition showed no significant change in comprehension performance from pretest to posttest, whereas the control condition exhibited a decline. Our results further indicate that mind-wandering is reflected in aberrantly slow but not fast sentence reading times. Results of the present study suggest that in addition to the previously established positive effects of metacognitive training on reading performance, the training is also effective in reducing task-unrelated thoughts during reading.