Due to the impact of the outbreak, many meetings have chosen to be held online. Video call technology is widely used in large-scale conference remote online occasions. However, due to perspective and distance issues, the current speaker of the meeting cannot always stay in the center of the video window. If multiple pan-tilt switching windows are used, the burden of the meeting will be increased. If a single wide-angle camera is used, either the picture will be blurred after the angle of view is narrowed, or a clear image of the participants cannot be obtained during panoramic shooting. In order to maintain the stability and fluency of the conference process, an intelligent PTZ system based on sound source positioning can be used, so that the current speaker in the conference can always be in the center of the video window, so that the conference can proceed smoothly and reduce the burden of the conference. The system is mainly composed of five-element cross microphone array, FPGA core controller and steering gear camera. The microphone array is used to collect the voice information of the speaker and transmit it to FPGA for calculation and positioning. FPGA is mainly used for calculation and control. After determining the position of the speaker, the amplitude of the steering gear to be rotated is calculated, and finally the steering gear is controlled to turn to the speaker. The steering gear is mainly controlled by FPGA and is used to provide stable video image signals for the conference. The microphone array of this system uses five-element cross silicon microphone, and the five-element cross microphone positioning system has the advantages of small size, fast operation speed and low cost, and can quickly obtain the sound source coordinates and control the steering gear rotation by using the powerful computing power of FPGA. Basically, it realizes the function of accurately positioning the speaker position, quickly rotating the steering gear, and providing stable video images in real time.