Driven by the digital revolution in second language (L2) classrooms and broadened views on literacy, digital multimodal composing (DMC) has gained a robust growth of interests in the past decade. To illuminate the empirical landscape of this budding field, this paper provides a substantive and methodological review of 60 empirical L2 studies on DMC in tertiary settings from 2005 to 2020. Each study was coded for characteristics regarding research context, methodology, study setup features, analyses performed, and the methodological practices. Our findings suggest that with a combination of flourishing qualitative observational research and emerging interventionist studies, this domain features (a) great attention to DMC in language classes, (b) an underrepresentation of non-English classrooms, (c) a preference for video projects over visual projects, and (d)immense efforts in exploring the potential of DMC from learners' perspectives. Also, using a broad range of data sources, this domain relies on analytic frameworks distinct from traditional monomodal L2 writing. Methodologically, we found a strength in the use of two validity strategies (i.e., a thick description and triangulation) and less-than-ideal practices in reporting reliability and learners' proficiency level. We conclude the paper with a number of empirically grounded recommendations for future research efforts.