As artificial intelligence and digital technologies are widely used in writing instruction, classrooms are adopting technology-supported feedback. This study investigates changes in learners’ self-regulated learning in a ChatGPT-based feedback English writing course. The research question for this study was: How does self-regulated learning change across motivational, cognitive, and behavioral domains before and after the course? The study was conducted in the English writing classes during the fall semester of 2025. A total of 48 students participated over a 15-week period. Data were collected through preand post questionnaires, in-depth interviews, and classroom observations. Quantitative data were analyzed using the Wilcoxon matched pairs signed-rank test while qualitative data were analyzed to support and explain the quantitative results. The major finding was that statistically significant changes were observed only in learners’ self-efficacy within the motivational domain over the course. The finding suggests that ChatGPT-based feedback is effective in supporting learners’ motivational regulation, particularly self-efficacy. However, to promote broader development of self-regulated learning, AI-based feedback may need to be accompanied by instructional support that guides learners’ monitoring, decision-making, and sustained engagement.