This study investigated how multimedia glossing affects incidental vocabulary learning from a listening task on mobile devices. A total of 118 English language learners were asked to listen to a story with 25 glossed target words on their mobile phones. In order to examine the effects of different types of glossing, participants were divided into four groups with access to four glosses during their listening: L1 textual, L2 textual, L1 textual and pictorial, and L2 textual and pictorial. Two vocabulary tests (i.e. definition-supply test and meaning-recognition test) were administrated immediately after treatment and two weeks later to measure vocabulary gain for target words. The results indicated that participants who had access to L1 textual and pictorial glosses had significantly higher vocabulary gains than other conditions, especially in meaning-recall word knowledge. Finally, a detailed discussion of the findings was provided to explain the results based on the theoretical framework of the study.