After the earthquake, buildings were damaged, and the environmental situation in the disaster area was complex. When using the machine for search and rescue, the traditional visual machine cannot accurately search and rescue the disaster victims, so it is necessary to cooperate with other perception methods to supplement the surrounding information. Sound is the appropriate way to obtain information. So, based on the algorithm of sound field imaging, we designed an intelligent robot that can obtain and track the surrounding sound information, and intend to apply it in the post-disaster search and rescue scenes. This design can identify and judge the direction of the surrounding sound signals, and drive the robot to move toward that direction. Meanwhile, it uses the OpenMV to upload the video streams via the WiFi module, thus realizing the function of real-time observation of the image information returned by the robot on other network devices. In the case of an unstable environment after the earthquake, this design can be applied to detect and conduct search and rescue operations in unstable areas.