<p>This study investigates the relationship between Cone Penetration Test (CPT)-based liquefaction severity indices and observed building damage following the 2011 Great East Japan Earthquake in the reclaimed Hinode district of Itako City, Japan. Relating liquefaction severity indices to observed building damage remains challenging because the influence of different CPT-based liquefaction triggering procedures has not been systematically evaluated against observed structural performance. Within a GIS framework, 143 CPT soundings were spatially correlated with damage records from 2,159 buildings classified into small, moderate, and large angular distortion categories. Liquefaction Potential Index (LPI) and Liquefaction Severity Number (LSN) were calculated using the Robertson and Wride (1998), Boulanger and Idriss (2014), and Architectural Institute of Japan (2001) procedures. The dataset enabled a systematic comparison of how different triggering procedures influence severity indices and their relationship with observed building damage. Both LPI and LSN increased systematically with increasing building damage severity. RW1998 and BI2014 provided better discrimination between small and moderate damage levels, whereas AIJ2001 produced consistently higher index values with greater overlap among damage categories. The separation between moderate and large damage was limited by the small number of heavily damaged buildings. Receiver Operating Characteristic (ROC) analyses indicated moderate predictive capability (AUC = 0.50–0.71). Although neither index alone provided high discriminative capability, both provide practical, locally calibratable tools for liquefaction-risk screening and post-earthquake damage assessment. These findings demonstrate that triggering methodology significantly influences index–damage relationships and should therefore be carefully considered when applying liquefaction severity indices in engineering practice.</p>