Abstract: To optimize the energy consumption of multiple mobile robots for parking and avoiding collisions, we propose a rolling Time Window and Binary Tree preorder traversal (TW & BT) fusion algorithm using a dynamic collision avoidance-integrated mechanism for multi-mobile robots under energy consumption optimization. First, we obtain the optimal initial paths using the improved A^* algorithm. Using collision-triggered events of the optimal initial paths, we decompose the whole operation time axis into several time windows. This process minimizes the total energy consumption for parking and avoided collisions. Since one time window corresponds to one sub-problem, we obtain the optimal collision avoidance strategy using the binary tree preorder traversal method. Compared with the dynamic priority conflict resolution strategy, the simulation experimental results show that the TW & BT fusion algorithm has higher robustness, and energy consumption decreases by over 33.1% in the approximate CPU time.