This study analyses the written production of EFL learners with the purpose of describing how these learners interact with multimodal texts (MT) in writing tasks whose aims are the development of writing skills, visual literacy and genre awareness. Three areas of research served as the basis for the study: multimodality, multimedia learning and genre-based teaching. Studies of visual language (Kress & Van Leeuwen, 1996; Rose, 2001) provided the elements for a model of analysis to be used in language teaching classes. The multimedia learning theory, conceived by Mayer (2000), grounded the examination of the verbal/visual combination in the teaching of writing. The idea of genre as social action (Miller, 1994; Bazerman, 1997, 2005) guided the activities designed for genre awareness. The data include the texts written by three adult learners in tasks with specific pedagogical purposes and different types of images/MT: a magazine cover with a photograph, an advertisement with a photograph and a textbook illustration. It also includes the recorded material of two of the oral analyses in class and questionnaires completed by the students after the writing task. The analysis looked for evidence in the written texts and in the learners’ discourse concerning: (1) interaction with the image/MT: perception, dependence and use of the represented elements, through the analysis of the topic framework of the text (Brown & Yule, 1983) and the learners’ responses to the questionnaires; (2) learning opportunities focusing on lexical improvement and genre awareness; (3) tacit knowledge of the genre and its influence, as well as the influence of the macro context on the learners’ interaction with the image/MT. Some of the findings suggest that (1) dependence on the image/MT is closely related to the learner’s writing ability; (2) the genre selected for the activity strongly influences the learners’ responses and their use of the visual and verbal resources of the image/MT, and (3) multimodal text analysis can enrich the foreign language class, motivating learners to exercise their visual literacy. The knowledge of multimodal learning can contribute to the design of materials for the teaching of writing.