The emergence of big data predictive policing has greatly improved the capacity of police to solve cases,prevent and crack down on crimes,but it also brings about certain risks that need to be regulated.Extraterritorial predictive policing has gone through three stages,namely 1.0,2.0 and 3.0.The prediction capacity has been continuously improved,and the operation mechanism has been gradually optimized.The development of big data predictive policing faces two risks:first,it is hard to ensure the accuracy,completeness,and timeliness of data;second,it is difficult to avoid or correct errors and discrimination in the algorithm.Big data predictive policing shall be regulated in the following three ways:first,the procedures for data collection and processing should be standardized to ensure the quality of data;second,the algorithm review mechanism should be established to supervise and evaluate the accuracy and risks of the algorithm;third,the establishment and use of predictive policing systems should be strictly regulated to ensure the protection of civil rights.