A region of interest (ROI) video request processing method and apparatus, where the method includes obtaining an ROI video request sent by a receiving device; selecting a first proximity ROI video from proximity ROI videos, where the proximity ROI videos are generated ROI videos that include at least a partial region in a target region indicated by the ROI video request; and sending the first proximity ROI video to the receiving device. According to the embodiments, the ROI video request can be met without consuming an encoder resource to break through limitation caused by an encoder resource amount to a quantity of ROI video requests that can be simultaneously met by the sending device.