The aim of this study is to compare effectiveness and efficiency of least-to-most prompting and video modeling for teaching pretend play skills to children with autism spectrum disorder. The adapted alternating treatment model, a single-subject design, was used in the study. Three students, one girl and two boys, between the ages of 5-6 participated in the study. The effectiveness results of the study showed that there is no marked difference between least-to-most prompting and video modeling for teaching pretend play skills to children with autism spectrum disorder in terms of acquisition, maintenance, and generalization. However, when these two teaching processes are compared in terms of efficiency parameters, it was observed that teaching with least-to-most prompting is more efficient in comparison to video modeling for two subjects. The social validity findings of the study showed that the mothers of the subjects and the graduate students who studied and had already taken certain courses for their master's degree in the Applied Behavior Analysis Program expressed positive opinions about the study. Findings obtained from the study were discussed and suggestions were given for further studies.