As K-12 classrooms in the U.S. grow in linguistic diversity, there is a need to develop and evaluate the effects of instructional approaches that build knowledge of academic vocabulary for multilingual learners. While multimedia vocabulary instruction may represent a promising avenue for improving word learning for linguistically diverse classrooms of upper-elementary students, there is a dearth of empirical evidence on this topic, with no research on multimedia vocabulary instruction having been conducted with linguistically diverse populations in rural contexts. This randomized control trial evaluated the effects of a multimedia vocabulary instructional approach (CAP-S), relative to business-as-usual (BAU) vocabulary instruction on researcher-developed and standardized measures of word knowledge for a linguistically diverse sample of Grade 5 students learning in rural schools (n = 418). The intervention took place over 8 weeks; CAP-S and BAU teachers addressed the same target set of academic words. Hierarchical linear modeling indicated statistically significant and meaningful effects on a custom measure of word learning, as well as small but statistically significant effects on a standardized assessment of word knowledge. On the custom vocabulary measure, ELs benefited more than non-ELs from CAP-S instruction, relative to their counterparts in the BAU condition. Findings suggest the potential advantages of employing efficient, feasible multimedia vocabulary approaches to serve diverse, multilingual student groups in rural, upper-elementary classrooms.