Flood-induced local scour is a significant factor in causing damage to bridges. However, the uncertainty of hydrological information and the impact of multi-layer soil greatly affect the accuracy of predicting scour depth for bridge foundations. To reasonably assess the safety performance of bridge structures during their operational lifespan, this paper develops a prediction method for scouring depth, a risk assessment curve, and a time-dependent vulnerability analysis model for bridge piles based on probabilistic hydrological information. Using a real bridge engineering example, the paper discusses the influence of probabilistic flood scour on the time-dependent bearing capacity, load effect, and failure probability of the bridge pile foundation. The results demonstrate that the proposed analysis method can effectively predict the relationship between scour depth and the service time of the bridge due to random stream-flow. The predicted scour depth of the pile foundation initially increases rapidly, then slows down with the service time of the bridge, indicating a significant increase in scour depth in the first 50 years, followed by a slowdown. However, the uncertainty in hydrological information can significantly impact the prediction results, especially in multi-layer soil scouring scenarios. The variation regularity of the mean and standard deviation of the time-dependent bearing capacity of pile foundation is consistent with the time-dependent regularity of scour depth. Moreover, the horizontal bearing capacity of the pile foundation is particularly sensitive to the scour depth during the same service time. The average vulnerability and standard deviation of the pile foundation under probabilistic erosion conditions increase with the duration of service year by year. [ABSTRACT FROM AUTHOR]