<p>Beam failure due to shear is particularly important in deteriorating structures, highlighting the need for prediction models to assess the impact of reinforcement corrosion on shear strength. Hence, this study focused on developing a robust predictive model to predict the shear strength of corrosion damaged beam. Machine learning approach was employed to capture the dominating material interactions and to accommodate a broader range of shear strength parameters. Parametric investigation was employed to systematically examine the relative contribution of each parameter on the overall behavior of shear strength. Various performance measures revealed that the proposed hybrid neuro-swarm model was found to be superior to other existing shear prediction equations. Using the causal inference procedure and parametric analysis, the effects and relative importance of each parameter on shear strength were explored. The parametric investigation offered clear and intuitive analysis that facilitated the identification of underlying patterns and correlations, which would have been challenging to recognize or interpret accurately through numerical analysis alone. The neuro-swarm model can be used to reasonably predict the shear strength of corroded beams and is vital in assessing the durability of deteriorating reinforced concrete structures. The results of the model can be used as a reliable baseline information for implementing appropriate remedial measures to prolong the service life of deteriorating structures. This is study is the first to employ machine learning techniques, particularly neuro-swarm algorithm, to develop a robust model to predict the shear strength of corroded reinforced concrete beam.</p>