Purpose: This randomized controlled trial evaluated the effectiveness of a play-based pragmatic language intervention for children with autism. Methods: A sample of 71 children with autism were randomized to an intervention-first group (n = 28 analyzed) or waitlist-first (n = 34 analyzed) group. Children attended 10, weekly clinic play-sessions with a typically developing peer, and parents mediated practice components at home. The Pragmatics Observational Measure (POM-2) and the Social Emotional Evaluation (SEE) evaluated pragmatics before, after and 3-months following the intervention. Results: POM-2 gains were greatest for intervention-first participants (p = 0.031, d = 0.57). Treatment effects were maintained at 3-month follow-up (p < 0.001–0.05, d = 0.49–0.64). POM-2 scores were not significantly different in the clinic and home settings at follow-up. Conclusion: Findings support the combination of play, peer-mediation, video-feedback and parent training to enhance pragmatic language in children with autism.