The high-frequency signal injection-type maximum torque per ampere (MTPA) algorithm is usually employed to control the operation of interior permanent magnet synchronous motors (IPMSMs). The MTPA algorithm exhibits good dynamic performance and anti-interference ability. However, due to the injection of a high-frequency current signal, problems such as torque ripple and additional loss are encountered. Therefore, in this paper, a virtual signal injection control (VSIC) method that does not require actual injection is proposed for solving the aforementioned problems while yielding good performance. However, in the control process of the proposed method, the d-axis inductance parameter affects the accuracy of the torque information, resulting in errors in the system. To solve this problem, an online identification algorithm of model reference adaptive systems (MRAS) based on the Popov super stability theory as the basis for the design of the adaptive law is proposed in this paper. The d-axis inductance parameter of the motor is obtained in real-time and then introduced into the control system by using the VSIC method. Finally, VSIC-type MTPA control based on inductance identification is realized. The proposed algorithm does not depend on the design parameters of the motor and exhibits good dynamic response and anti-interference performance.