Objective: to establish a methodology for designing an automatic control system (ACS) and a corresponding state observer for induction traction motors (ITM), specifically to guarantee high-precision torque regulation characterized by defined accuracy, rapid dynamic response, and the absence of overshoot. Methods: the foundational step in achieving the stated objective lies in selecting an optimal control strategy for the ITM that guarantees the necessary performance metrics. Consequently, this paper investigates the core principles of control and structural configuration of the motor’s ACS. This study advocates for the implementation of Direct Torque Control (DTC). By regulating the flux-linkage vector via the voltage vector, this method offers distinct advantages, including a predictable harmonic spectrum for both currents and voltages, alongside the requisite dynamic behavior. Furthermore, to mitigate the accumulation of integration errors and ensure the robust functionality of the state observer, the research proposes utilizing numerical solutions for differential equations based on the Euler and Adams — Moulton methods. Results: to validate the proposed design methodology for the ACS and the ITM state observer, comprehensive mathematical modelling of the traction electric drive system has been conducted. The simulation data confirms that an ACS based on the principles of direct torque and voltage control delivers superior dynamic regulation capabilities. Notably, the system achieves rapid torque control while maintaining a comparatively low switching frequency for the inverter’s power semiconductor components. This efficiency is largely attributed to the removal of stator current con- trol loops, which subsequently allows for optimal utilization of wheel-rail adhesion conditions. Practical sig- nificance: the study underscores the critical importance of selecting a suitable ITM control strategy to secure high precision, speed, and the overall control quality. Employing the second-order Adams-Moulton method effectively neutralizes error accumulation with minimal additional computational load. This mathematical adjustment is instrumental in ensuring the efficient and reliable operation of the state observer in practical applications.