Abstract: Crest cracking is one of the most common damage types for high core rockfill dams. Cracking risk of dam crest is closely related to the duration of abnormal deformation state. In this paper,a methodology for dynamic risk evaluation and early warning of crest cracking for high earth-rockfill dams is proposed. The abnormal state of the dam crest is identified by the settlement inclination method. The exponential distribution is adopted to represent the probability distribution of the duration TAS of abnormal state before crest cracking. The dam crest cracking risk assessment model is constructed to realize real-time analysis of dam crest cracking risk and probability-based cracking time estimation. Bayesian estimation and updating of probability distribution of the parameter λ in the risk model,according to observed durations of abnormal state before crest cracking. Validity of the methodology is illustrated and verified by the case study for a high core rockfill dam.Based on the dam crest cracking observed at the time of the high water level over the years,the parameter λ of risk model is updated to gradually reduce the dispersion of the probability distribution,and the confidence interval of the dam crest cracking probability estimation is also reduced. The dam cracking time is predicted by the statistical analysis of the corresponding cracking probability when the dam crest is cracked,and the early warning of crest cracking is issued based on the lower limit of the confidence interval of the cracking time estimation.The result shows that the method can predict the cracking time of the dam crest accurately and issues an early warning of crest cracking more than 20 days in advance.