Abstract: This study investigates innovative approaches to AI-enhanced experimental teaching within the context of marine-themed integrated physical chemistry experiments. The research demonstrates the effective utilization of knowledge graphs for systematic learning and interdisciplinary knowledge integration. By incorporating DeepSeek and Python-based open-source models, the study significantly enhances intelligent data processing capabilities, thereby promoting the advancement of smart experimental teaching methodologies.

Key words: Physical chemistry experiments, Artificial intelligence, Marine-themed, Knowledge graph, DeepSeek, Python