Abstract: In order to achieve high-performance control of induction motor, a sensorless direct torque control system is proposed to address the shortcomings of traditional speed identifiers. In this system, based on the mathematical model of induction motor, after certain transformations, a speed identifier is designed using the easily detectable stator voltage and current of the motor, as well as a neural network based on the BP algorithm, thereby realizing sensorless closed-loop control. The designed system was simulated, and the results show that the speed control system has good steady-state performance and dynamic performance. The design is reasonable and effective. Keywords: Direct torque control; Sensorless; Neural network; Induction motor 0 Introduction Direct torque control (DTC) of induction motor is a high-performance AC speed control technology. It abandons the decoupling idea in vector control and directly calculates and controls the flux linkage and torque of AC motor in the stator coordinate system using the space vector analysis method based on the detected stator voltage and current [1]. In this high-performance speed control system, the introduction of a speed closed loop is essential. Speed ​​detection devices typically employ tachogenerators or photoelectric encoders. The installation of these speed sensors not only increases the cost of the control system, affects detection accuracy and control performance, and reduces reliability, but also limits their application range due to their unsuitability for harsh environments such as humidity and high electromagnetic noise interference that affect signal transmission. This paper introduces neural network control theory, employing neural networks and model reference adaptation to establish a speed identifier, which is then applied to a direct torque control system. This achieves sensorless closed-loop control, increasing the system's adaptive speed regulation capability while meeting the requirements of fast response and high steady-state accuracy. Sensorless Direct Torque Control System for Induction Motors