Studying the underwater behavioral characteristics of shrimp is important for shrimp farming. In order to achieve efficient, rapid and accurate shrimp identification and behavior observation, a detection and behavior analysis method of Penaeus vannamei in aquaculture ponds based on active identification sonar (DIDSON) is proposed. Eight observation points were set up in the shrimp pond, each with a distance of 40 m between them, and the two banks were symmetrically distributed. Since the traditional way of setting up brackets for fixedpoint observation will consume a lot of manpower and material resources, a set of DIDSON sonar observation system based on intelligent double-hull unmanned boats is designed, that is, the use of unmanned ships to carry active identification sonar (DIDSON) can realize continuous fixed-point observation of 8 observation points in a short period of time ( observation time of each observation point is 3 min, interval time is 7 min). In this experiment, a shrimp identification and counting model was constructed based on the underwater acoustic data processing software ECHOVIEW, and the image data collected were identified and extracted, and the preliminary behavioral characteristics of Penaeus vannamei in the aquaculture pond were obtained by target recognition, target extraction, flux analysis and cruise direction analysis. The extreme values of the number of shrimps observed in the eight observation points were 251 and 208, respectively, and the average value was 234, indicating that the distribution of Penaeus vannamei shrimp in the breeding pond was relatively uniform. The flux interval of shrimp swarm per unit time was [108, 131] / (min·m 2 ), the mean value was 122 / (min· m 2 ), the standard deviation was 6 / ( min·m 2 ), and the flux change was small. The proportion of positive swimming ponds of shrimp groups at different points was greater than 85%, and the highest could reach 92. 57%. The results show that this method can observe the regular swimming behavior of shrimp groups in aquaculture ponds, which effectively solves the problem of observation of shrimp groups in aquaculture ponds compared with traditional underwater visual observation and passive sonar investigation, and provides scientific data support for the formulation of more efficient shrimp pond feeding and management schemes. [ABSTRACT FROM AUTHOR]