In the last decade, several researchers have explored the use of traditional feature selection techniques for identifying software metrics that effectively identify aging-related bugs. This study proposes an enhancement to existing research by using explainable AI to select metrics that are most likely to cause software aging. To this end, we have introduced an automated approach using SHAP (SHapley Additive exPlanations) that has been validated on seven open-source software datasets. Furthermore, the study conducts a comparative analysis of our proposed approach with two traditional feature selection techniques (Relief & Gain Ratio). The results evaluated using AUC and F1-score favor the use of SHAP for feature selection, as it provides an average AUC score of 0.766 and an F1 score of 0.897. Moreover, the proposed framework allows interpretability of the selected features that can aid software practitioners in preventing the introduction of aging-related bugs.