ABSTRACT Cables have usually served as critical and vulnerable structural components in long span cable-supported bridges. Cable inspections revealed that corrosion, fatigue or coupled corrosion-fatigue were the ones of the main failure mechanisms. This paper proposed a stochastic modelling method for three-dimensional (3-D) corrosion pits of high-strength bridge wires, which can be applied to rapid fatigue life evaluation according to mass loss caused by surface corrosion pits of bridge wires nondestructively. High-strength steel wire specimens dismantled from the cable-stayed bridge served for 15 years were scanned to obtain the original surface corrosion data. The spatial position coordinate of corrosion pits were considered as random variable and can be well fitted by uniform distribution. While the number of corrosion pits can be fitted with generalized extreme value (GEV) distribution. The uniform corrosion depth du, which can be equivalent to mass loss rate, was calculated as the input corrosion parameter for 3-D corrosion pit modelling. The maximum pitting depth dmax for the steel wire was found to be associated with du. The geometric parameters for individual corrosion pits were recognized as pit depth d, depth-to-width ratio d/b, and aspect ratio b/a, which were fitted with different probability distributions. What follows is 3-D spatial corrosion pits simulation based on the individual corrosion parameter that were sampled and combined from the corresponding probabilistic distributions. Hereafter, fatigue life evaluation of corroded wires was conducted based on equivalent surface defect method and compared with the experimental results, verifying the effectiveness of the proposed modelling approaches.