Introduction: Studies have shown that many music learners experience music performance anxiety (MPA). Without appropriate intervention, MPA can erode performers’ confidence and diminish learning efficiency. To date, existing research has predominantly relied on cognitive behavioral therapy. This study developed and evaluated an intervention model that integrates artificial intelligence (AI) with e-learning tools to alleviate MPA and enhance performance-related outcomes.Methods: A purposive sample of 64 undergraduate music majors was recruited and randomly assigned to an experimental group (n = 32) or waitlist control group (n = 32). During the eight-week intervention, the experimental group received AI-assisted positive feedback in addition to real-time visual feedback through digital practice tools, while the control group received standard instruction. Data were collected on music performance anxiety, flow experience, emotional states, psychological distress, expert-rated performance quality and signs of performance anxiety. Data were analyzed using repeated-measures ANOVA and paired-sample t-tests.Results: The results showed a significant reduction in MPA with a notable group × time interaction (F(1, 62) = 8.79, p < .001, η² = .12). Flow experiences also improved, as demonstrated by the significant interaction effect (F(1, 62) = 4.30, p < .05, η² = .07). In the experimental group, negative emotions and signs of performance anxiety decreased significantly.Discussion: The findings indicated that AI-assisted positive feedback can effectively reduce MPA and promote flow experiences among music students. The study further introduces a multi-dimensional intervention mechanism—“real-time feedback → skill improvement → enhanced flow → increased self-confidence”providing new insights into MPA interventions and strategies for enhancing musical performance.