Convolutional neural network (CNN) and machine deep learning are used to analyze the required static parameters of the growth environment of shrimp farming. Using the parameters, the intelligent monitoring system analyzes the needs of different shrimp larvae with different growth environments. Image recognition technology and the collection of sensor data are used to control various dynamic parameters, such as feeding, growth, movement, and accident warning. The edge computing of the proposed system can improve the efficiency of shrimp farming production and avoid accidental injuries caused by emergency accidents.