Multi-path deep learning framework on discrete pressure points to predict velocity field of pump-jet propulsor

•A fast and accurate deep learning framework for predicting the velocity field of pump-jet propulsor has been investigated based on discrete pressure points.•Once the deep learning model is trained, the velocity field of pump-jet propulsor can be obtained only through discrete pressure points, such as discrete pressure points obtained by a pressure sensor.•The accuracy and generalization ability of model are intuitively analyzed by statistical parameters such as peak signal to noise ratio (PSNR) and linear correlation coefficient.