Micro-nanorobots swarm based on magnetic nanoparticles has many applications in the medical field such as minimally invasive surgery and drug delivery. Many studies have aimed to analyze the motion behavior of nanoparticle swarm through physical modeling. However, these physical models primarily focus on different experimental phenomena and struggle to achieve a general and real-time simulation of multiple particles. The absence of universally applicable physical modeling for nanoparticle swarm significantly increases the training cost and efficiency of using AI to control the swarm and limits its expansion in virtual reality applications. We developed a physics-based simulation of a nanoparticle swarm based on Unity, which can simulate the motion response of swarm in different magnetic fields in real-time. Initially, we analyzed the physical forces acting on particles and established functions to describe the main forces based on physical parameters. We set active region for each particle, disregarding insignificant long-range forces, greatly reducing computational complexity. By comparing simulation results with experimental outcomes under different types of magnetic field inputs, we observed good consistency between the simulation and the reality.