To address the limitations of conventional remote-controlled miniature vehicle platforms, including a single human–machine interaction interface and cumbersome operation, this study proposes and develops an intelligent autonomous driving car system based on voice interaction. The system employs the STM32F103C8T6 microcontroller as the core processing unit and deeply integrates the LU-ASR01 embedded voice recognition engine to achieve efficient non-contact command control. Digitalized voice feature codes are transmitted to the main controller through UART asynchronous serial communication, and PWM (Pulse Width Modulation) technology is further utilized to drive the L293D dual H-bridge motor module, thereby enabling omnidirectional kinematic control of the vehicle, including forward motion, reverse motion, steering, braking, and multi-level speed regulation. Experimental results demonstrate that the proposed system exhibits a high command recognition accuracy, an end-to-end response latency strictly within 2 s, and excellent dynamic stability and robustness under complex testing conditions. This work provides a reliable design paradigm and theoretical basis for the engineering implementation of voice recognition technology in miniature intelligent mobile robots and embedded control systems.