Controlled studies examining canine therapy in autism spectrum disorder are scarce. This study examined the effectiveness of a “Dog Training Intervention” on adaptive skills, autism severity, and anxiety using a controlled crossover design. Seventy-three participants diagnosed with autism spectrum disorder ( Mage = 4:10 ± 1:0) were divided into two groups that received the dog training intervention during half of the school year in addition to standard-of-care interventions. The dog training intervention, in which the children were taught how to interact with and train dogs, was given twice weekly for 4 months within autism spectrum disorder–specific special education school. Those receiving the dog training intervention first showed significantly increased adaptive social and communication skills compared to the controls, and the gains were maintained after the dog training intervention. Belonging to the first dog training intervention group, higher pre-intervention adaptive skills, higher baseline cognitive ability, and less severe autism severity predicted better adaptive social and communication skills. The controls improved in adaptive skills only during their receipt of dog training intervention after crossover. The positive impact on social communication skills suggests that dog training may serve as an effective model for establishing social interaction. Dog training intervention appears to be an effective adjunct treatment to interventions provided in special education schools for children with autism spectrum disorder. Lay abstract There is some evidence that using therapy dogs for children with autism spectrum disorder generally results in improved social communication skills and reduced behavioral problems. However, well-controlled studies that examine its effectiveness are scarce. This study examined the effectiveness of a “Dog Training Intervention.” The study included 73 participants diagnosed with autism spectrum disorder (61 males, 12 females) with age range of 2:10–7:6 years ( M = 4:10 ± 1:0) who attend autism spectrum disorder–specific special education schools. The study population was divided into two groups. Each group received the dog training intervention during one part of the school year (first half or second half) in addition to the standard interventions provided by the special education school settings. The dog training intervention was given twice weekly for 4 months within the school setting. The group that received the dog training intervention first showed a significant increase in adaptive social and communication skills in comparison to the second group that did not receive the intervention in this period. This improvement was maintained after the dog training intervention. The second group, which received intervention at the second half of the year, showed improvement in communication and socialization adaptive skills only during the period in which they received the dog training intervention. The positive impact on social communication adaptive skills of the dog training intervention among young children with autism spectrum disorder suggests that dogs may serve as an effective model for establishing social interaction. Dog training intervention appears to be an effective adjunct treatment to the interventions provided in special education schools for young children with autism spectrum disorder.