This study examines the potential of educational media to provide preschool-aged dual language learners (DLLs) with vocabulary in a new language. Drawing from dual-coding theory, the current study investigated how three distinct instructional contexts with varying degrees of incidental-intentional vocabulary exposure on screen might facilitate second language (L2) vocabulary development. Instructional contexts included participatory contexts that engaged viewer attention and elicited viewer response surrounding a vocabulary word (least incidental, most intentional), expository contexts that provided explicit visual-auditory vocabulary scaffolds (somewhat incidental, somewhat intentional), and narrative contexts that embedded vocabulary words in conversations within a storyline (most incidental, least intentional). The study used a within-subjects design with 50 preschool-aged DLLs. Children watched nine 2-minute video clips, followed by vocabulary knowledge assessments. Findings indicate that instructional contexts were differentially facilitative in helping DLLs identify words in a new language, F(1, 47) = 11.003, p = 0.002. Moreover, L2 proficiency moderated the influence of instructional contexts on vocabulary identification but not word meaning. Results suggest that media programs with relatively intentional exposure to vocabulary words on the incidental-intentional continuum may scaffold L2 word learning better than incidental exposures. Implications on the critical role of incidental vocabulary exposure in media learning environments for preschoolers are discussed.