For the controllable excitation magnetic levitation linear synchronous motor (CEMLLSM), the conventional iterative learning control (ILC) has low precision, large chattering and poor resistance to external disturbances. In order to improve the tracking accuracy, a variable gain adaptive ILC algorithm based on extended state observer (ESO) is designed. Firstly, the working principle and mathematical model of the CEMLLSM are studied. Secondly, a variable gain adaptive iterative learning controller based on ESO is designed. An exponential variable gain is introduced to the fixed gain part of the controller, and an adaptive iterative term is added to iteratively learn the unknown parameters in the control law, which reduces the chattering and error of the system, and speeds up the system convergence speed. By adding ESO, the external disturbance of system can be observed, and the control quantity is compensated to enhance the anti-disturbance performance of the system. Finally, the control system is simulated and analyzed with MATLAB, and the results show that the algorithm can successfully reduce tracking error and has a good suppression impact on disturbance.