Background: There is a growing body of literature investigating the efficacy of music interventions for children with autism spectrum disorder (ASD); however, little empirical research has been conducted into the use of musical elements to facilitate language learning. Methods: This crossover-design study compared the responses of 22 children with ASD (M age = 5.88 years) to sung and spoken instructions embedded into a computer-based communication intervention designed to teach receptive labelling. Results: There was no significant difference between the sung and spoken conditions. Following both conditions, there was a significant increase in receptive labelling skills; skills were generalised and were maintained at follow-up. A difference in group performance was found. Conclusion: Further research is required to investigate child characteristics that may impact on children's performance using this approach.