In 2020, the state issued a ten-year ban on fishing in the Yangtze River basin, which has put forward more stringent requirements for public security officers to combat illegal fishing crimes. At present, there are problems such as low control efficiency, limited monitoring scope and no direction for arrests in combating illegal fishing activities in the Yangtze River basin. In this paper, we use police data as samples, extract the characteristics of criminal personnel using the TF-IDF algorithm, then use these characteristics as weights, use the combination of coefficient of variation plus hierarchical analysis algorithm to predict the key suspect, and finally use the trajectory carving model to carve the front, middle and back ends of the suspect’s crime execution to establish an early warning model of illegal fishing crimes in the Yangtze River basin. It provides investigative ideas for the detection of illegal fishing of aquatic products, curbs the spread of illegally caught aquatic products, improves regulatory measures and forms a normal and long-term mechanism for prevention and control and governance.