Fire flame identification research, as one of the key technologies of intelligent fire protection, plays an important role in intelligent fire protection equipment. Combined with the research results of flame recognition algorithms at home and abroad in recent years, this paper makes a review and analysis of the research status of fire flame identification algorithms. Focus on the comparative analysis of related algorithms in the field of flame recognition research based on shallow learning algorithms such as SVM, decision tree, wavelet neural network, and deep learning algorithms based on R-CNN series algorithms, YOLO series algorithms, and SSD. At the end of this paper, the future development direction and trend of flame recognition algorithm are prospected, in order to provide relevant theoretical and experimental research reference for flame recognition research through this paper.