In this paper, we study the collision-free routing of a multi-robot system to complete given tasks in the shortest time. In a robotic assembly unit, several stations work serially and in parallel. In a station, multiple robots share the same workspace and face the challenge of minimizing the cycle time and avoiding collisions at the same time. For this problem, we propose a new mathematical model that is the so-called timespace network (TSN) model. The TSN model can map the robot location constraints into the routing planning framework, leading a mixed integer programming problem. By solving this mixed integer programming problem, the collision-free path of multiple robots can be obtained. Finally, The simulation results illustrate the proposed TSN model can obtain the collision-free route of the multi-robot system.