This paper investigated the relationship between the morphology of corrosion pits and the stress concentration factor. Several images of corroded wire with different corrosion durations were obtained based on corrosion-accelerated test under the coupling effect of alternating load and salt-spray environment, then the morphology and size of pits were identified by the maximum entropy threshold segmentation method. A finite-element model of high-strength wire with corrosion pits was constructed and analyzed to explore the stress concentration factor close to the corrosion pits. Subsequently, a Bayesian prediction model of the stress concentration effect on wires with corrosion pits was proposed. In conclusion, the morphology of corrosion pits was described as “ellipsoidal”, with its length and width obeying the lognormal distribution and its depth obeying the Weibull distribution. Bayesian analysis indicated that the stress concentration factors around a corrosion pit were affected by the length–width and width-depth ratios of the pit. The stress concentration effect close to corrosion pits with narrow or deep appearance was more dramatic than that of other pits.