A method of identifying a level of doneness of food includes: obtaining an initial food detection image; obtaining a current food detection image; performing a differential fusion operation on the initial food detection image and the current food detection image to obtain a fused food detection image; and inputting the fused food detection image into a predetermined neural network model to obtain the level of doneness of food output by the predetermined neural network model. According to the method, the differential image fusion is performed to reduce an influence on determining the level of doneness of food caused by the initial food detection image, and costs of the device may be reduced.