The unbalanced force on the active magnetic bearing (AMB) rotor system is the main reason of system vibrations, which always have periodic repetitive characteristics. However, the closed-loop feedback control system (CFCS) is difficult to achieve effective control of these repetitive problems. In this paper, we firstly analyzed the synchronous vibration by taking AMB rotor mass imbalance as an example. Then, we proposed a model-free control (MFC) method based on the Newton-type ILC algorithm, and the key of this MFC is to use a partial derivative (P-D) of the output with respect to the input. The simplicity of this algorithm lies in that and the P-D is calculated by only using the input and output (I/O) data of the system and then used to adjust the ILC gain. Subsequently, we proposed a parallel plug-in control scheme based on existing AMB control system to suppress synchronous vibration. Finally, we carried out the simulation and experiment to research the control method mentioned in this paper. The results show that the MFC has a good control effect on AMB synchronous vibration. Notably, this MFC for synchronous vibration has the advantages of being without system model, simple design and good portability, and can provide theoretical and experimental basis for the application of AMB in high precision and high speed fields.INDEX TERMS Active magnetic bearing, synchronous vibration, model-free, iterative learning control parallel control scheme.