This paper presents a prognostic technique for the Remaining Useful Life (RUL) of a ball bearing. A state-space model based on Paris" law is applied to estimate the damage progression. The hidden damage state and the time to failure are computed by using noisy measurements obtained from a accelerometer. In order to identify the model parameters, a series of run-to-failure tests for ball bearings has been carried out. A Particle Filter approach is used to predict the demage progression rate and update the degradation state based on the most recent measurement. The future state and the RUL are predicted by the assumption that the previous state is not different from the unobserved measurement. The developed method was validated within the predicted uncertainty by comparing the prognostic results and test data.