This study proposes an adaptive graph algorithm for collision-free motion planning of articulated robots in dynamic environments. For this purpose, deformations of the configuration space were analyzed according to the changes of the workspace using various simulations. Subsequently, we adopted the principles of gas motion dynamics in our adaptation algorithm to address the issue of the deformation of the configuration space. The proposed algorithm has an adaptation mechanism based on expansive repulsion and sensory repulsion, and it can be performed to provide the entire adaptation using distributed processing. The simulation results confirmed that the proposed method allows the adaptation of the roadmap graph to changes of the configuration space.