This study is to focus on the marine cage culture industrial problem of optimal feeding, monitoring cage environment, and labor-saving. The surface and underwater precision feeding monitoring system was tested on July 24-25, August 15-16, and November 11-12. The objective is to carry out image recognition monitoring and management for the fish feeding and fish school of the cage culture. Based on the test results of the water spray detection algorithm on July 25th, the statistical data of large splash: in the first stage of feeding were significantly different from large splash: in the latter stage of feeding. It can be speculated that after feeding, most of the fish have gradually become full or the appetite for feeding has decreased, resulting in a decrease in the degree of splashing. Secondly, in the intelligent hydrological environment monitoring system of marine aquaculture nets, take July 25 as an example: average temperature (28.16 ° C), specific conductivity (52.09 mS / cm), total dissolved solids (33.86 ppm), and salinity ( 34.2).Finally, in the foreign applied technology of automatic laborsaving tools, the physical mechanical cleaning method is relatively safe, reliable and low pollution. The existing physical mechanical cleaning methods mainly include drum type net washing machines, integrated screen washing machines, high-pressure water washing methods and mesh cleaning machines. The study is expected to save manpower, reduce the cost of aquaculture production, and further achieve accurate feeding management to maintain the optimal fish growth and cage culture environment.