Artificial intelligence (AI) technology, exemplified by DeepSeek, is reshaping the form and pattern of education and teaching in an unprecedented manner, pioneering a new paradigm in the field of physics education. This article integrates various AI tools to conduct innovative teaching practices for complex physical laws and models through dynamic visualization. Leveraging the Matlab and Python programming platforms, we carefully select “dynamic simulation of a highly adjustable snake pendulum” and “construction of atomic lattice imaging using scanning tunneling microscopy” as core teaching cases, striving to establish an innovative teaching model that encompasses “physical modeling-simulation replication-parameter tuning-autonomous exploration”. This teaching model effectively overcomes the physical limitations faced by traditional experimental setups, successfully constructs dynamic physical processes involving multiple parameters, and designs models with dynamically adjustable parameters through programming. Students undergo training from “observation” to “interpretation”, enabling them to gain a deeper understanding of abstract physical concepts, and their scientific research interests and innovation abilities are simultaneously enhanced. Furthermore, by guiding students to explore and replicate academic research frontiers, we greatly stimulate their interest in learning, exercise their systematic thinking, and thus promote students to actively engage in curiosity-driven learning and research.

In the information age, resources are abundant and knowledge is rapidly updating, but the basic physical concepts and theoretical principles remain immutable. How to transfer mundane physical formulas and complicated physical laws into intuitive and dynamic demonstrations during the classroom teaching process, as well as igniting students' enthusiasm to stimulate their in-depth thought have become a problem and challenge faced by Collage Physics. In this paper, we take “the composition of multiple vibrations and travelling waves” as an example, using Matlab software to build a visual physics demonstration platform, dynamically demonstrating the physical processes of beats and standing waves. By slightly adjusting parameters such as amplitude, frequency, and wave speed, the evolution laws and visual analysis of amplitude, phase, and energy ofconventional standing waves after the superposition of beat effects are explored. Further, an interactive operation interface is introduced to display the visualization of complex motion and sound output of virtual demonstration experiments. These pedagogical endeavors can greatly enhance students' spatial imagination with expand threedimensional scientific thinking. Simultaneously, such teaching attempts encourage and guide students to search for related literatures and independent practice, which further promotes their engagements in freely exploring a broader array of physical phenomena in the natural world.