Аннотация:The article considers the development of predictive policing in Russia through the creation of software, based on the use of artificial intelligence (AI) to identify serial killers. Forensic modelling in crime investigation, in particular modern digital twin technology is analyzed. The system of the digital twin is trained on the basis of a set of mathematical models of various level of complexity and specified by results of full-scale experiments. Existing approaches to solving the serial killer portrait problem are investigated. Digital twins in conjunction with machine learning can predict the behavior of the object under study in the future, based on statistical data and accelerate the work of the investigator in the investigation of serial crimes.