A dynamic personal space model and its parameter generation method are proposed to achieve socially-aware robot navigation. Firstly, a basic social cost model that can flexibly represent the concentric circle shape, the concentric ellipse shape, the egg-like shape, and the asymmetric shape are established. Secondly, a dynamic model based on the relative speed of pedestrian and robot and robot mood regulation is established. Finally, the parameter generation method of the proposed model is introduced. The effectiveness of the model is verified through simulation and real-life experiments.