Beyond effectiveness of the instructional model of Navigation of Social Engagement (NOSE model; Chou, 2020) to enhance problem solving skills of students with autism spectrum disorders (ASD), this article addresses moderating factors to provide a more in-depth secondary analysis of data within this intervention. While previous findings have established the effectiveness of the instructional model, the current study further examined the differential efficacy of the intervention by analyzing the moderating factors-inclusiveness (i.e., education placement, interaction with peers, time spent with peers without identified disabilities), cognitive functioning (i.e., above-average, average, mildly impaired, and severely impaired), and social support (i.e., support from family, peers, and teachers). The analytical results of testing the model with three-way interaction terms showed that cognitive functioning and time spent with peers without disabilities were the two positive significant moderators of the intervention effect, indicating that (a) the self-directed social problem-solving model had the optimal effect when students with ASD were included with their peers, and (b) that students with mildly impaired cognitive functioning had the greatest improvement in social problem-solving compared to students with above-average, average, and severely impaired cognitive functioning. This analysis of moderating factors provided additional information on the NOSE model for promoting self-directed social problem-solving for students with ASD.