This study explores the possibilities of using digitally produced multimodal language learning histories (MLLHs) to understand the experiences of learning a second language (L2) of a group of university students in Japan (N = 21). The study observes that learners' MLLHs are texts in which the visual elements of language, place, person, learning resource, and self-analysis of learning process are represented visually. It also identifies 4 patterns in the ways the MLLHs are constructed by focusing on a certain type of visual element. Person-oriented MLLHs focus on L2 learning as emotional experience. In resource-oriented MLLHs, learning is perceived to occur because of engaging with language via favorite media. In analysis-oriented MLLHs, learning is regarded as a matter of going up and down the scale of linguistic measures. Place-oriented MLLHs emphasize being on site at an L2-speaking region as being the key in the era of globalization and mobility. These patterns reflect learners' beliefs on L2 learning as viewed from a contextual perspective. This study further argues that a multimodal approach can have a significant influence on making visible learners' subjective perspectives and beliefs as lenses through which they frame their learning experiences, as well as visualizing the dynamics and individualities of learning.