In view of the complex and low identification accuracy of load and mutual inductance identification methods in magnetically coupled wireless power transfer (MCR-WPT) systems, a method has been proposed for simultaneous identification of load and mutual inductance in MCR-WPT systems. Firstly, with the influence of high-order harmonics in the system taken into consideration, a steady-state circuit model of the LCC-S topology system is established and verified, thus obtaining mathematical descriptions of the output current, mutual inductance, and load of the primary inverter. Secondly, an improved particle swarm optimization algorithm is introduced so as to construct a fitness function based on the error between the theoretical and actual values of the output current of the primary inverter, with the parameter identification of the system transformed into an issue of optimization, thereby obtaining the identification values of the system load and mutual inductance. Finally, the simulation results show that the load and mutual inductance errors identified by the proposed method are kept below 1.20%, which is very close to the actual values, verifying the feasibility of the proposed method. This proposed approach is applicable to both resonant and non-resonant MCR-WPT systems without compensation network constraints, thus enabling precise parameter identification without primary-secondary communication or auxiliary circuits.