In terms of the diagnosis of inter-turn short circuit (ITSC) faults of dual-three-phase permanent magnet synchronous motors (DTP-PMSMs), the conventional methods are easily affected by factors such as fault saliency and feature extraction frequency, and the fault index fluctuates with the change of operating conditions, making it difficult to maintain the diagnostic accuracy. Therefore, a closed-loop diagnosis method using model reference adaptive system (MRAS) observer is proposed in the essay. Firstly, the ITSC models are established according to the structure diagram of MRAS system. And the harmonic current is obtained by the space vector decomposition (VSD). Then, based on the residual signal, the fault characteristics of the fundamental component of the short-circuit current are extracted, and the adaptive law is designed through the piecewise PI controller, so as to implement closed-loop adjustment of the parameters to be evaluated in the adjustable model. Finally, a phase-shifting 30° asymmetric six-phase PMSM is chosen as the object of simulation analysis, and the results show that the proposed diagnostic strategy has better diagnostic accuracy and adaptability than the traditional open-loop diagnostic method. Compared with the traditional open-loop fault diagnosis system, the diagnosis error of the proposed strategy is reduced by 20-25% in the working range where the speed and torque fluctuate ±by 40% compared with the rated operating conditions.