Remaining Useful Life (RUL) prediction is important for the prognosis and the maintenance of expensive systems. High reliability is an indispensable requirement for advanced systems. Correspondingly, reliability practitioners attempt to investigate the system behaviour in order to mitigate risk as much as possible. The study of the behaviour is insured by collecting the real data of degradation, which is considered as a challenged task. Based on the retrieved data, one can propose a good stochastic process to model the system degradation. In this study, the Variance Gamma process is proposed to model the degradation of the water tank pump. The estimation of the data parameters and their fitting to the model is considered. In addition, the study of the First hitting time (FT) and its distribution is treated. Finally, a good prognostic is conducted based on the FT results.