Children with sleep problems present serious management problems to their parents. Such children are also more likely to have additional problems, behavioural disturbance being particularly common. This randomized controlled trial of behavioural interventions for the children's sleep problems was conducted to explore the efficacy and mechanisms of treatment in children with the most extreme forms of problems: severe learning disabilities, severe sleep problems and severe daytime challenging behaviour. Fifteen index families received behavioural advice for the child's sleep problem and compared with 15 matched controls who received no such advice. Repeat assessments of the children's and mothers' sleep were made by parental report as well as actometry. Objective changes in the children's sleep quality and quantity were not seen after treatment. However, mothers in the treatment group reported improvements in the children's sleep problems and had an increased sleeping time themselves following treatment. The results indicate that sleep problems can be successfully treated in this group of children, although the mechanisms of treatment may not be as direct as supposed. This has implications for understanding of sleep problems in children with learning disabilities and also for clinical practice, when considering ways of offering help to these highly 'challenged' families.