This study's objective was ascertain, through the use of self-efficacy measures as mediator variables, how well the group counseling model applied to modern puppet media improved the vocabulary mastery of students with special needs. The study used a quasi-experimental research design. The independent variables were the group counseling model with modern puppet media, self-efficacy as a moderator variable, and vocabulary mastery as a dependent variable. The effectiveness testing of independent variables on dependent variables and their interaction with moderator variables was designed using a pretest-posttest control group design. Utilizing the cluster random sampling technique, a sample of up to 78 students from the experimental group and 83 students from the control group were obtained. Self-efficacy was measured using a standardized self-efficacy scale, while vocabulary mastery was measured using a vocabulary mastery test. To test the hypothesis, a two-way ANOVA was used, and further testing was carried out using the Scheffe' method. The data analysis results indicate a noteworthy distinction in vocabulary mastering between students receiving a modern puppet-based group counseling model and students receiving a traditional learning model; there is a significant difference in vocabulary mastery between pupils with low and high levels of self-efficacy; and there is an interaction between the group counseling model with modern puppet media and self-efficacy on students' vocabulary mastery.