One of the main issues in aquaculture is the high cost of the lost food when the fish are fed. Furthermore, this wasted food is deposited in the seabed and generates an environmental impact on the surrounding area. It causes failure to uphold aquaculture legislation. The authors propose a feeding control system based on groups of sensors that take the appropriate decisions when the fish are fed in marine fish farms. This system mixes data obtained from a group of sensors and performs an exhaustive control on the food offered to the fish, thus reducing costs. This will avoid wasting large quantities of food and an economic saving will be achieved. To develop the proposed system, the authors used the behaviour of the movements of the fishes and some parameters from the water. These movements are monitored strategically by a group of underwater transducers placed along the length, width and depth of the cage and, through data fusion, they provide the information needed in the system to continue feeding the fish or not.