An ultrasonic wave-based voice signal transmission system, where the system includes an ultrasonic modulator, a beamforming controller, an ultrasonic transducer array, and a user detector. The ultrasonic modulator is configured to modulate a voice signal onto an ultrasonic band and output the modulated voice signal to the beamforming controller. The user detector is configured to detect a user and output a detection result of the user to the beamforming controller. The beamforming controller is configured to control according to the detection result from the user detector, a phase and amplitude of the modulated voice signal to obtain an electrical signal pointing to the user, and output, to the ultrasonic transducer array, the electrical signal pointing to the user. Therefore, call convenience can be improved for the user.