Post-earthquake evaluations of structures are crucial for determining usability or the need for repairs and retrofitting. This study employs Artificial Neural Networks (ANN) for damage detection in reinforced concrete (RC) frame structures by analysing their dynamic characteristics, such as natural frequency, and mode shape. Using modal analysis in ETABS software, we obtain these characteristics for a three-storied 2D RC building frame. Damage is simulated by altering structural rigidity, creating a dataset for ANN training. The trained ANNs predict damage based on deviations in natural frequencies, revealing a strong correlation between damage levels and dynamic characteristics. We discuss the application of these correlations for evaluating the health status of structures post-earthquake, underscoring the effectiveness of periodic vibration monitoring as a non-destructive method. Further research is needed to optimize ANN-based SHM and extend its applicability to other structural types.