This project proposes a fundamentally different and innovative fatigue prognosis methodology based on a small time scale formulation is proposed for the real-time concurrent structural damage prognosis. The proposed novel damage model overcomes the inherent difficulties in existing fatigue theories. One of the most important benefits is that concurrent fatigue analysis across multiple spatial and temporal scales becomes feasible. Pervasive prognosis capability is addressed in this study, from material level up to structure level. Rigorous validation of model hypotheses and prediction will be performed using state-of-the-art experimental techniques, such as in-situ fatigue testing under scanning electron microscopy combined with digital image analysis. A special focus in the proposed study is on the systematic uncertainty modeling through multilevel computational simulations. Advanced reliability methods, Bayesian statistics and information theory are proposed to capture the stochastic nature of fatigue damage accumulation.