Abstract: In view of the non-linear parameter variation of the hybrid excited axial field permanent magnet (HE-AFPM) motor caused by the vector coupling effect between the excitation and armature flux, the end leakage flux effect, and etc. in multiple operating modes, an adaptive model predictive current control method is proposed in this paper. The machine topology and parameter variation characteristics of the HE-AFPM motor are analyzed in detail. The multi-vector model predictive current control is introduced, and the sensitivity of the HE-AFPM motor parameters to the control system is theoretically derived. On this basis, the multi-vector predictive model is adopted to the adjustable model of the adaptive control, which reduces the complexity of the control algorithm and the computational load of the control system. In addition, according to Popov's hyper-stability theory, three-inequalities stability criterion is introduced to design the adaptive law of each parameter. In this way, the parameters including flux linkage, self-inductance, and mutual inductance are effectively identified. And then, the accuracy of the control model and the current tracking performance are improved. Finally, the effectiveness of the proposed adaptive model predictive current control is verified by the experiments.