Abstract: Due to the randomness of material properties and external loading, the predicted remaining fatigue life under variable amplitude loading often has a very great dispersion. Thereby, to quantify the uncertainty of the remaining fatigue life arising from material properties and external loading, this paper proposes a remaining fatigue life prediction method under two-level loading. This proposed method mainly consists of four steps. Firstly, considering the randomness of external loading, a nonlinear damage accumulation model based on continuum damage mechanics is applied. Secondly, considering the randomness of material properties, three-parameter S-N curve under different reliability is adopted. Thirdly, in order to screen the model parameters that contribute greatly to the randomness of the predicted remaining fatigue life, a global sensitivity analysis based on Sobol method is introduced. Finally, the K-L divergence method is utilized to quantify the differences between the predicted and the probability distribution of the experimental remaining fatigue life. In addition, the effectiveness of the proposed method is verified through the fatigue test data of hot-rolled 16 Mn alloy steel.