This study aims to examine the effectiveness of teaching naming emotional facial expression via video modeling to children with autism. Teaching the naming of emotions (happy, sad, scared, disgusted, surprised, feeling physical pain, and bored) was made by creating situations that lead to the emergence of facial expressions to children participating in the study. A multiple probe design with probe trails across behaviors was used and replicated across subjects. Four subjects diagnosed with autism (ages 4,5, and 6) participated in the study. Ten people participated in the research study: eight people for the video modeling display and two for displaying the emotional facial expressions in the teaching and generalization sessions. The findings of the study show that teaching the naming of emotional facial expressions to children with autism is effective, and after the end of training the children can still maintain this ability. In addition, subjects can generalize their acquired skills from a variety of simulation situations created by different materials in different environments, both with persons who participated in this model and those who did not. Social validity findings show that the children's mothers and teachers and post-graduate students have positive opinions about the study.