Undergraduate students face significant psychological challenges, including diminished academic self-efficacy (ASE) and heightened fear of negative evaluation (FNE), which impact academic performance and university adaptation. This study evaluated the effectiveness of AI chatbot-enhanced e-counseling platforms in addressing these critical concerns among university students. A quasi-experimental design was employed with 66 undergraduate students randomly assigned to experimental (n=34) or control (n=32) groups. The experimental group received access to an AI-enhanced e-counseling platform incorporating cognitive-behavioral therapy techniques and personalized support over eight weeks, while the control group received standard university services. Participants completed assessments at pre-intervention, post-intervention, and four-week follow-up using validated measures of ASE and FNE. Results revealed significant improvements in the experimental group, with ASE scores more than doubling from baseline (M=22.85 to M=46.23) and sustained at follow-up (M=45.88). FNE significantly decreased from pre-intervention to post-intervention and follow-up. Large effect sizes were observed for ASE (η²p=.598) and medium effects for FNE (η²p=.156). The AI-enhanced e-counseling platform demonstrated substantial and durable effectiveness in improving psychological outcomes among university students, suggesting promising applications for scalable university mental health interventions.