Nowadays, the Physics-Informed Neural Network (PINN)-based computational mechanics is being rapidly developed, although it is currently still in its infant stage. It has been becoming a game-changer for computer modelling and simulation for engineering and science. This talk will first overview the most recent developments of PINN for computational mechanics. The main challenges in PINN based computational mechanics are discussed. Second, the recent research in the speaker’s group on PINN will also be reported in this talk, including the new loss function approach for PINN and applications in different fields. It has proven that physics-informed machine learning will be a new powerful tool for computational mechanics.