Several experimental studies have found that rolling element bearing's service life can be typically divided into three phases: normal operation, gradual degradation, and accelerated degradation. Although extensive studies have been carried out on bearing life prediction, most studies focus only on life prediction in the accelerated degradation phase. Although Oil Debris Monitors (ODMs) have excellent potential in bearing condition monitoring, only a handful of studies explore its potential for Remaining Useful Life (RUL) prediction. Therefore, in this thesis, an intelligent bearing health management system is presented that primarily uses the information from the ODMs. The system contains two major modules namely, fault phase diagnosis and RUL prediction. The data-driven diagnosis module uses multi-sensor data along with three novel degradation indicators based on wear debris characteristic information. The RUL prediction module is a model-based tool where a novel physics-based multi-phase degradation model is utilized in a novel enhanced adaptive Particle Filter.