Virtual reality technology has been widely applied in simulation systems and visualization applications in various fields. Unity3D has realistic 3D rendering effects and powerful cross platform features, providing a powerful visual editor and rich graphical interfaces, making it very suitable for developing virtual experimental systems. This article is based on Unity3D to design a virtual experimental system. Firstly, this article investigates the foundation of Unity3D technology, including the composition of Unity3D and the application of MVC in Unty3D; then, collision detection algorithms were studied. In this paper, the particle swarm optimization algorithm from artificial intelligence was introduced into the field of collision detection. A mathematical model was constructed and simulation experiments were conducted; finally, the design and implementation work was completed, and a virtual experiment development environment, virtual experiment scene construction, and virtual interaction function implementation were designed, laying the foundation for system development.