The ability to track and follow a specific person is a key prerequisite of service mobile robots in indoor environments. A novel following method is introduced, which uses the sizes changes of torso in image as the following control parameters and the fuzzy PID controller is employed to design the following controller. The sizes of torso in images are obtained with mobile robot vision tracking method which integrates the color features of upper body clothes region matching method with the Unscented Particle Filter (UPF) tracking method. In order to reliably follow a specific person, the parameters of the PID controller are adjusted using PSO (Particle Swarm Optimization). Various experiments are carried out and the results demonstrate the robustness and reliability in person-following.