In the application of swarm robot systems, basic behaviors such as aggregation, diffusion, and obstacle avoidance are fundamental research issues in the study of swarm robot system control, and are often the prerequisite or important basic work for completing other tasks. Inspired by the movement phenomenon of algae systems, we propose a uniform diffusion controller for distributed multi-robot systems, which consists of only two actions based on binary-state environmental information obtained by each robot's sensors; no Additional memory and communication are required. We performed simulations using 100 epuck robots through the enki simulator to validate the effectiveness of the controller and analyze the diffusion process for different numbers of robots. Experiments with 20 SwarmBang robots are also conducted to demonstrate the proposed controller. In addition, the diffusion method we propose has the characteristics of low requirements for robots, fast diffusion, and uniform distribution, which provides an alternative solution for solving the diffusion problem of ultra-large-scale robotic systems.