While mathematics education is key to the post-school outcomes of students with disabilities, it has received less attention in research and practice compared with other aspects of educating this population. Skill maintenance is particularly crucial in mathematics because students build upon prior knowledge across grade levels. They also need to be able to apply mathematical skills in everyday life. Hence, this study utilized a multiple probe across-participants single-case experimental design to evaluate the effectiveness of the virtual--representational--abstract (VRA) with overlearning instructional sequence in teaching multiplication and supporting its maintenance among three students with disabilities. For each student, a functional relation existed between the VRA with overlearning instructional sequence and accuracy of solving multiplication problems. Students also maintained the skill up to 8 weeks after the intervention.