This study presents a probability-based surface deterioration evaluation framework to evaluate the cracking damage of bridge pylons subjected to repetitive thermal stresses caused by daily temperature variations. First, a temperature prediction model for massive concrete bridge members was utilized, incorporating the effects of convection and radiation. Subsequently, the thermal loads applied to the bridge pylon were computed using the finite element (FE) method. A modified linear elastic fracture mechanics (LEFM) crack model combined with random field theory for predicting crack propagation in concrete was employed. A probability-based surface damage grade method for a bridge pylon was proposed. Moreover, the deterioration state evaluation was updated by incorporating Bayesian inference, which involved using in situ inspection results (i.e. spatial distribution associated with surface crack lengths). The methodology was implemented using an illustrative example that considers a bridge pylon. The results provide useful information for aiding decision-making strategies concerning the maintenance and repair of massive concrete bridge components.