Two functional assessments (interview and direct observation) were used with three children with autism to identify the functions maintaining mealtime behaviour including acceptance, mouth clean, refusal, and other disruptive behaviours such as crying and pushing the spoon. Based on results of the functional assessments it was hypothesized that appropriate and disruptive mealtime behaviour was maintained by different contingencies. A non-concurrent multiple baseline design across participants was utilized to validate the effectiveness of the intervention. Intervention for all participants included presentation of food on a spoon for 30 s unless acceptance occurred. Acceptance resulted in putative reinforcement. The meal ended after 20 presentations. For all participants, acceptance and mouth cleans increased while disruptive behaviour decreased, and effects were maintained at follow-up.