This study investigated whether a modified scoring method was useful for examining the ability of children with social communication disorder (CwSCD) to understand non-literal language and use mental state responses on the Happé Strange Stories (HSS) task. CwSCD and a control group of children with typical language development (CwTLD) completed 10 of the original HSS. CwSCD scored significantly lower on the HSS task than did CwTLD and were much less likely to produce mental state responses. There was a high level of inter-rater reliability (Weighted Kappa = 0.907) across data from both groups. HSS performance and language ability correlated significantly for CwSCD. A regression model with age, nonverbal intelligence, receptive and expressive language as predictors explained 55.2% of the variance in HSS ability for CwSCD. The results suggest that the HSS have potential to be used as a clinical assessment to investigate high-level language and ability to infer intent in CwSCD.