Background: This study compared the effectiveness of computer-assisted instruction to that of one-on-one tutoring for teaching people with mild and moderate cognitive disabilities when both training methods are designed to take account of the specific mental deficits most commonly found in cognitive disability populations. Method: Fifteen participants (age 22-71) received either computer-assisted instruction or one-on-one tutoring in three content domains that were of functional and daily relevance to them: behavioural limits, rights and responsibilities (two modules) and alphabetical sorting. Learning was assessed by means of a series of pretests and four learning cycle post-tests. Both instructional conditions maintained time-on-task and teaching material equivalence, and both incorporated a set of best-practices and empirically supported teaching techniques designed to address attentional deficits, stimulus processing inefficiencies and cognitive load limitations. Results: Strong evidence of learning was found in both instructional method conditions. Moreover, in all content domains the two methods yielded approximately equivalent rates of learning and learning attainment. Conclusions: These findings offer tentative evidence that a repetitive, computer-assisted training program can produce learning outcomes in people with mild and moderate cognitive disabilities that are comparable to those achieved by high-quality one-on-one tutoring.