The increasing use of Artificial Intelligence (AI) in language learning has opened new opportunities for improving second language (L2) pronunciation skills. This study aims to evaluate the effectiveness of integrating artificial intelligence into second language pronunciation training by examining its impact on learners' outcomes through a mixed-methods approach. Quantitative results from 60 participants showed the AI group (M=4.89) significantly outperformed the control group (M=3.71) in post-tests (p=0.001), with an 85% vs 65% improvement rate. Qualitative interviews revealed three key findings: (1) real-time feedback enhanced engagement and self-confidence, (2) notable improvements in intonation and stress patterns, but (3) limitations in recognizing regional accents. The results demonstrate that AI-assisted pronunciation training offers clear advantages over traditional methods in enhancing L2 pronunciation. However, limitations in accent recognition highlight the need for further development of AI speech processing technologies. This study provides empirical support for integrating AI into language education, highlighting its potential and areas that require improvement.