This study aimed to examine incidental learning of multiword expressions (MWEs) from bilingual captioned videos among language minority students. The participants consisted of 403 language minority university students enrolled in the Preparatory Year Programme (PYP) in the Southern part of China. The study evaluated participants' vocabulary knowledge (VK) by examining both its breadth and depth, alongside their working memory (WM) capacity in tasks involving phonological and complex processing. Additionally, a test on meaning recognition of MWEs was administered. The results revealed that: (1) bilingual captioned videos facilitated the incidental learning of MWEs; (2) both VK and WM capacity were significant predictors of incidental learning of MWEs from bilingual captioning and; (3) VK mediated the relationship between WM and incidental learning of MWEs at the level of meaning recognition. These findings underscore the benefits of incorporating bilingual captioning in learning MWEs for language minority students. However, it is crucial to consider the students' prior VK and WM capacity.