The parameter identification of a power electronic converter can improve the control and operational effect of systems. However, in traditional parameter identification methods, it is difficult to identify multiple groups of parameters at the same time, and the accuracy of the identification results is low. To solve this problem, a parameter identification method of three-phase inverters based on a digital twin is proposed. First, a digital twin three-phase inverter is constructed, including the establishment of the mathematical model of the main circuit of the three-phase inverter using a fourth-order Runge-Kutta method and the establishment of a discrete model of the controller. Then, the circuit parameters of the digital twin inverter are updated and optimized by an adaptive particle swarm optimization algorithm until the corresponding circuit parameters of the digital twin inverter and the physical inverter are the same. Finally, the effectiveness of the proposed parameter identification method is verified by simulation and experiment. The results show that the inductances, parasitic resistances and switch tube internal resistances of the physical inverter can be quickly identified under both steady and dynamic conditions, and the relative errors of identification results are within 2%.