Mining industry is a key sector of many national economies, thus even a small crack in a mining site echoes nationwide. This strategic sector is subjected to special safety care in every aspect, including operational safety of engineering structures. The newest technologies are employed to diagnose and monitor the working structures and buildings. In this article we propose an innovative idea of combining GPS monitoring system with artificial neural network prognosing to build a prediction tool for displacement of mining shafts in operational conditions. The paper describes a training of a neural network system whose task is to prognose the displacements of the top of a mine shaft tower in direction in a selected period of time. The data used for the training come from the GPS monitoring of displacements of the top of the S 1.2 mining shaft tower. The tower is located in the mining works LW “Bogdanka”.