<p>To facilitate the development of a performance-based framework for liquefaction hazard assessment in a regional scale, this paper proposed probabilistic predictive models for estimating the permanent settlement of low-rise buildings with shallow foundations on liquefiable sites based on numerical simulation results and available case history database. The primary predictors of foundation settlement were identified as foundation bearing pressure, foundation embedment depth, depth from the bottom of the foundation to the middle of the first susceptible sand layer, foundation length-to-width ratio, as well as empirically computed liquefaction-induced free-field ground settlement. The cumulative absolute velocity was identified as the optimal intensity measure (IM), utilized to develop a single IM model. A second set of models using PGA and earthquake magnitude, which were easily assessed as the intensity measure, was developed. However, this model showed greater variability compared to the CAV-based model. The proposed models captured the key trends in permanent foundation settlement observed in available field case histories with simple functional form and easily assessable IMs. These results highlight the potential tradeoffs associated with model uncertainty and efficiency in the model application that should be considered in liquefaction hazard assessment. Moreover, the models provided a link between the liquefaction triggering analysis framework and the estimation of engineering demand parameters, enabling regional estimation of liquefaction consequences.</p>