Based on Skinner’s classification of verbal behavior, the mand is the first and most advantageous verbal operant to develop. Deficits in vocal mand repertoires are common in children with autism spectrum disorders (ASD) and can lead to decreased social interaction and increased problem behavior. The present investigation compared the effects of two mand training procedures on the acquisition and mastery of vocal mands by four young children with ASD using an adapted alternating treatment design. An in vivo mand training procedure, which involved the therapist contriving an establishing operation and delivering a vocal model, was compared with a video modeling procedure wherein the therapist presented a video model prior to contriving an establishing operation. Overall, participants acquired and mastered mands assigned to the video modeling condition faster than mands assigned to the in vivo mand training condition. The results demonstrate the potential efficacy of video modeling for teaching verbal operants to children with ASD.