People with mild intellectual disabilities (ID) often have difficulties understanding and expressing advanced linguistic concepts. Prior to the arrival of COVID-19, there were no research-backed models to guide virtual interventions with this population. The research presented here pursues three objectives: to determine if the vocabulary of people with mild ID improves following face-to-face and virtual learning interventions, to understand the differences between these two interventions, and to assess participants’ perceptions of both modalities. The sample consists of 10 adults with mild ID who participated in 10 face-to-face learning sessions and 10 virtual learning sessions. An ad hoc designed vocabulary assessment instrument and semi-structured interviews are used, in addition to statistical software SPSS and Atlas.ti. The results show improvements in vocabulary in both modalities, although no significant differences were found, suggesting that both in-person and virtual modalities are effective for vocabulary learning in people with ID. However, there were exceptions in some categories such as gardening and baking. Likewise, participants expressed satisfaction and enthusiasm toward both interventions, but showed a preference for face-to-face sessions.