In the operation cycle of reconfigurable manufacturing system, the system needs to be continuously reconstructed to adapt to the changes of market demand. Aiming at the problem of when to reconstruct the reconfigurable manufacturing system, combined with digital twinning, a prediction method of reconfiguration time point of reconfigurable manufacturing system is proposed. Firstly, the factors affecting the processing capability and function of reconfigurable manufacturing system are analyzed quantitatively, the optimization objective is to minimize the part completion time and reconfiguration cost, and the multi-objective cuckoo algorithm is used to solve it; Then, the digital twin model of reconfigurable machine tool and reconfigurable production line is constructed, so that the state of reconfigurable manufacturing system can be truly mapped in virtual space, and the simulation and prediction framework of reconfigurable manufacturing system is constructed; Secondly, the initial scheme configuration is generated according to the production demand, the set threshold is optimized and simulated, and the simulation sample data is analyzed to obtain the appropriate reconstruction time; Finally, an engineering example is given to verify the feasibility and effectiveness of the method.