This study focuses on Vietnamese traditional music genres such as Ca Tru, Don Ca Tai Tu, and Cheo, which face the risk of fading due to diminishing interest among younger generations. To address this, we propose VietHarmony, an AI-powered framework designed to preserve and revitalize these traditions. VietHarmony combines machine learning models with Borda voting to integrate outputs from various CNN architectures, including DenseNet169, MobileNet, and ResNet-50, achieving a remarkable F1-score of 98.43%. A key contribution is our novel FAMBO (Fusion of Audio classification Models using BOrda) algorithm, enhancing classification accuracy. We also introduce a \(\mathcal{E}\mathcal{L}_\bot \) lightweight ontology for extracting and presenting contextual information. Delivered via a user-friendly web platform, VietHarmony provides a comprehensive solution for the preservation and revitalization of Vietnam’s cultural heritage.