Emotion recognition has been identified as critical in the development of social communication and interaction with individuals having autism spectrum disorder (ASD). However, little is known about the effects of the intervention and moderator factors with quantitative research. To carry out an analysis of effect size on the intervention emotion recognition of ASD, and to propose an optimal intervention plan, a meta-analysis was conducted across 22 randomized controlled trial studies involving 817 individuals with ASD, with the aim of analyzing the effectiveness of moderator variables. The findings of this study showed that the effect size of emotion recognition intervention for ASD reached a medium level (d = 0.69), and the effect size of the technology-based intervention (d = 0.90) was significantly greater than that of the non-technology-based intervention (d = 0.37). Though technology-based intervention may be a promising method on emotion recognition for ASD individuals, however, more well-controlled studies are needed.