This study employed an explanatory mixed-methods design to examine how teacher-generated emergent feedback could affect EFL learners’ writing performance across high and low writing self-efficacy levels. It included 60 Iranian upper-intermediate EFL learners who received teacher-generated emergent feedback during a ten-week writing intervention. Quantitative data were collected through pre-test and post-test academic writing tasks, and learners. Given that the quantitative data were not fully normal, non-parametric analysis was employed. Quantitative analyses indicated a significant improvement in the writing performance of high self-efficacy group from pretest to posttest following exposure to teacher-generated emergent feedback. Low self-efficacy group, however, did not demonstrate a statistically significant pretest-posttest gain, although some learners reported more confidence when the feedback became more explicit. Between-group comparisons indicated significant differences at pre-test, post-test, and gain-score levels, showing that learners’ sense of efficacy was related to different developmental patterns. Qualitative data obtained from semi-structured interviews revealed that learners with high and low self-efficacy engaged differently with emergent feedback. High self-efficacy learners reported greater independence and strategic use of implicit feedback, whereas low self-efficacy learners relied more on explicit guidance and emotional support, particularly in early feedback stages. Taken together, the findings suggest that teacher-generated emergent feedback can support writing development across self-efficacy levels, while learners’ efficacy beliefs shape how feedback is experienced and utilized. The study contributes to L2 writing research by integrating performance outcomes with learner perspectives and by re-conceptualizing feedback as an emergent, interaction-sensitive instructional practice.