Selecting two heavy pollution weather processes cover Hubei on 25 December 2015 and 18 January 2016, and using FY3A (B)/MERSI satellite data and weather observation, environmental monitoring data, this paper carries out the study of haze remote sensing identification by digitizing the image color processing technology and technology of reflectance threshold of visible and near infrared channels in Hubei area. The results show that haze can be distinguished by the color composition of red, green and blue channels. With full visible bands, the color of haze is mainly graywhite, slightly darker than the surrounding cloud area; with visible, infrared and near infrared bands, the color of haze is mainly purple and purple grey, and the cloud class recognition is more fine; and with enhanced display of visible and near infrared bands, the color of haze is mainly purple grey, which can be distinguished from the clear sky surface and cloud area, but easy to be confused with the thin cirrus composed of small particles, so it needs to be eliminated through the blackbody brightness temperature. With the help of TBB of infrared channel and the visible reflectance index, haze can be identified from the clear sky and thick cloud area, but difficult to distinguish from mist. The problem can be solved by adding the reflectance of near infrared channel which is sensitive to the effective particle radius. Finally the results are verified through the observation of ground weather phenomena.