The invention particularly relates to a model evaluation method based on deep counterfeit video detection. The method comprises the following steps: s100, importing a model to be scored; s200, randomly selecting a specified number of samples from a video sample library to obtain a sample set, wherein videos and authenticity marks of the videos are stored in the video sample library; s300, inputting the sample set into a to-be-scored model for authenticity identification; s400, comparing the identification result with the actual authenticity of the sample to obtain a measured value P1 of a specified scoring index of the to-be-scored model; and s500, performing automatic scoring according to the measurement value P1 to obtain a score of the to-be-scored model, and outputting the score. According to the invention, a high-value detection model can be screened out and incorporated into a professional deep detection defense system, so that the deep detection defense system is constructed, and a network security defense line is created and consolidated; in addition, continuous forward development of a deep counterfeit detection model technology can be promoted, deep counterfeit model research and development personnel are attracted to participate in continuous improvement of a video sample library, and a virtuous circle is achieved.