An active random dynamic instruction generation system for resisting counterfeit face identification is characterized in that a face detection module is used for positioning a face and extracting features and storing the features in a database; the face detection module triggers a random dynamic instruction generation module and random dynamic instruction combination requiring a target face, eyes and a mouth is generated; after a random dynamic instruction is generated, a to-be-identified user does corresponding motions according to requirements with a system voice prompt, and the motions are subjected to one-to-one verification by a random dynamic instruction verification module; a target tracking module utilizes video target tracking technologies to carry out continuous tracking of areas of the face, the eyes and the mouth; and after the random dynamic instruction verification module completes identification, the face detection module extracts face features, and the face features are sent to a face identification module for completing final identity authentication and identification. By adding the active random dynamic instruction generation module and a random dynamic instruction interpretation module based on video tracking technologies, a face identification system is enabled to have an effective counterfeit-resisting function, and the system safety is greatly improved.