<p>This paper investigates the impact and assessment of regular and irregular buildings on structural damage using Artificial Neural Networks (ANN). The target audience is design engineers, building managers, building inspectors and public authorities. The following four parameters merit particular consideration: design irregularities and vertical irregularities in both directions in their planes, redundancy in the vertical shear resisting system, and the number of floors in the building. In this study, a novel approach based on an artificial neural network (ANN) was proposed for the assessment of the structural quality factor (Q_F) utilising a dataset of existing buildings from the Boumerdes post-earthquake, 2003 Algerian earthquake damage assessment forms. This approach is predicated on the implementation of a specific analytical expression, which facilitates a more accurate assessment of structural integrity. A comparison was conducted between the observed and the predicted damage, revealing a high level of concordance in terms of the location of the damage. The efficacy of this approach was evaluated by comparing ANN performance with that of other machine learning algorithms, including Support Vector Machines (SVM), Random Forests and Logistic-Reg. The proposed ANN attained 90,60% accuracy and an Area Under the Curve (AUC) of 0,984, thus confirming the ANN’s resilience to data set imbalance. In addition, the model displays a low mean calibration error (MCE) of 0.06, which is suggestive of optimal predictive reliability. The results obtained from the analysis indicate that the ANN model based on the quality factor provides a more effective and extensible solution for structural condition assessment in comparison to current advanced methods.</p>