In the production practice, it is important to detect the feeding behavior of fish for feeding control. Taking Cyprinuscarpiospeculari as an experimental object, this paper employs computer vision technology to detect feeding behavior by using shape and texture information of fish in the process of feeding.Firstly, the image is subtracted, grayed out, binarized and so on, and the image shape and texture information are obtained. Then, the feeding behavior of fish is classified by BP neural network. Compared with the single texture-based detection method, this method can not only analyze the interference of the unfavorable factors such as surface vibration, spray,and other adverse factors, but also consider the shape information of the image and improve the accuracy of the detection. The results show that the correct recognition rate of this method is 98.0%, which can be used to guide the precise feeding control in aquaculture.