Abstract: Anchor channels with channel bolts (AC-CBs), as steel-made anchoring connectors, are widely adopted in utility tunnels, tunnels, and other industrial structures for securing pipelines and equipment supports. In such service environments, AC-CBs may experience severe corrosion over their operational lifespans. Considering the stochastic nature of corrosion, this study conducts tensile simulations at two critical locations—the mid-span and the bolt centerline—for commonly used Q355 steel AC-CBs, evaluating the influence of different corrosion parameters on their mechanical performances. The research results show that the volume loss ratio Dov and the pit depth ratio η have the most significant weakening effect on the load-bearing capacity of AC-CBs. An extreme Gradient Boosting (XGBoost) algorithm based on an active learning strategy is developed, and SHapley Additive exPlanations (SHAP) analysis reveals that in the tensile capacity assessment of locally corroded AC-CBs, the weakening effect of pit size is negligible compared to that of Dov and η. Therefore, these two parameters must be considered in capacity calculations. Based on the above findings, a power-function degradation model is employed to establish a prediction formula and a simplified version for the reduction factor of the ultimate tensile capacity of AC-CBs. By incorporating a time-dependent corrosion function, a capacity evaluation method for AC-CBs in different geographical service regions is proposed.