For this study, 11 children with moderate to severe autism spectrum disorder (ASD) were given directives containing prepositions in three cue conditions: (a) spoken alone, (b) a short video clip along with spoken cues, and (c) a sequence of three graphic symbols accompanied by spoken cues. Participants followed directives significantly more accurately with the video clip than with spoken cues only, and significantly more accurately with spoken cues only relative to the sequence of graphic symbols. Results suggest that the short video clip along with spoken cues may be an optimal mode for enhancing learners' ability to follow directives containing prepositions. In addition, results reveal three statistically significant correlations between participants' preexisting skills and directive-following accuracy: a positive correlation between spoken preposition preassessment total score and accuracy in the spoken-alone condition; a positive correlation between spoken noun preassessment total score and accuracy in the video-clip condition; and a positive correlation between ASD severity and the need for repetition in the video-clip condition. Results also suggested that, for children with more severe ASD symptoms, the video clips require repetitions so that the relationships illustrated within it can gain more semantic salience. Implications for clinical practice and future research are discussed.