By constructing an algorithm model of vocal music acoustic feature extraction based on AI (Artificial intelligence) and optimizing teaching strategies, this article hopes to improve the efficiency and quality of vocal instruction. In order to achieve this goal, the vocal music sample data is collected and preprocessed, and the acoustic features are accurately extracted by deep learning method. Subsequently, this article designs individualized teaching strategies based on the results of AI acoustic analysis, including real-time acoustic feedback system and individualized training plan, to help students quickly identify and correct pronunciation problems and achieve targeted skills improvement. In the experimental verification stage, vocal music students were selected as experimental objects, and they were randomly divided into experimental group and control group, and the experimental period was set for several months. The final results show that the students in the experimental group are significantly better than those in the control group in key acoustic features such as pitch, resonance and vibrato, and the overall singing skills and expressive force have also been significantly improved. This shows that AI technology shows great potential and advantages in vocal music training, which can provide students with a more scientific, accurate and individualized learning experience.