Retrieval-based learning has been investigated in various populations. The present study examined retrieval-based learning in a sample of students at a special-needs school with educational tracks for learning and mental development. In addition, a comparison group of students at a regular school was examined. Learning conditions were manipulated within participants. In a first session, participants either received restudy cycles only, or they received alternating test and restudy cycles. A second session then comprised the opposite form of practice. In both sessions, a final test assessed memory after a short distractor phase. This procedure was the same in two experiments but with different kinds of item material. For both kinds (images and image-word pairs), a testing effect occurred, that is higher recall of tested items. These results show that lower cognitive ability or lack of experience with regularly being tested in school do not prevent students from benefiting from retrieval-based learning.