The tensile deformation of top-down cracks significantly influences pavement delamination. Accurately formulating the stress intensity factor (SIF) is crucial for assessing the stress/strain near crack tips. In the present study, a robust data mining method called M5 is employed to formulate the SIF induced by pure bending in a multilayer pavement. A comprehensive database is generated through two-dimensional finite element analysis to train the M5' model. Several key parameters pertaining to asphalt and base layers are incorporated into the formulation. A comparative analysis is conducted to evaluate the performance of the developed model against traditional regression methods. A comparison of the developed model with other existing methods confirms its superior performance. Additionally, a parametric study validates the robustness of the developed model in capturing fundamental concepts of fracture mechanics. Finally, a sensitivity analysis is performed to determine the most critical parameters in the development of a predictive model for SIF.|The mode I stress intensity factor of top-down cracks is computed for various conditions. The MNLR, MLR, and M5 methods are used to formulate the mode I SIF of top-down cracks. The fatigue life of the asphalt pavement is assessed using Paris' law. The thickness of the HMA layer is the most critical parameter in predicting SIF.