Many children with autism show very little interest in academic assignments and exhibit disruptive behavior when assignments are presented. Research indicates that incorporating specific motivational variables such as choice, interspersal of maintenance tasks, and natural reinforcers during intervention leads to improvements in core symptoms of autism and may possibly be effective in academic areas. Using a multiple baseline across children and behaviors design with four pre- and elementary school children with autism, we assessed whether the above variables could be incorporated into academic tasks to improve performance and interest. Results indicated that the intervention decreased the children's latency to begin academic tasks, improved their rate of performance and interest, and decreased their disruptive behavior. Theoretical and applied implications are discussed. (Contains 2 tables and 4 figures.)