Abstract:  Exercises and example problems are crucial means for consolidating physics learning. In current university physics instruction, exercises are often categorized by knowledge points, with relatively rigid problem-solving approaches. This practice can lead to rigid thinking in students analysis of physics problems, leaving them at a loss when encountering unfamiliar questions and fostering an aversion to physics due to perceived difficulty. Consequently, it hinders the development of students ability to transfer knowledge and innovate. Using mechanics as an example, this paper proposes a teaching design that interconnects multiple knowledge points through example problems and exercises. By integrating multiple knowledge points into a single problem, students can grasp physical concepts holistically, gain a deeper understanding of the inherent logical relationships between theoretical knowledge, broaden their analytical perspectives, promote deeper learning, and enhance their ability to solve practical problems.

Key words:  deep learning, knowledge connectivity, multiple solutions for one problem, problems of variable mass