The importance of soil-structure interaction can be seen in a large number of infrastructure’s project, where it causes a significant influence to project stability, safety and durability in geotechnical, seismic and structural engineering. Traditional methods of predicting structural response based on soil properties appears to rely on empirical or analytical methods; neither of which adequately capture the complexity of soil-structure interaction. But with the introduction of Machine Learning (ML), researchers can now handle massive data sets, and therefore it is now feasible to develop predictive models to enable greater accuracy in predictions. This study investigates the potential for ML algorithms to predict structural behavior from soil properties and applied loading. This research utilizes a two-phase methodology: first, an integrated data collection phase that gathers a collection of geotechnical and structural information. Second, a ML model is trained to categorize and classify liquefaction cases. The research demonstrates that the ML model capably predicts structural behavior by merging data collection processes with m algorithms. An appropriate ML model through its training process achieves the ability to predict soil-structure system behaviors under different loads. The model leads to steady predictions to which decision-makers can consult. Outcomes show how ML is efficient in forecasting geotechnical hazards and consequently design methodologies, providing the advantage for experts to predict main behavior and adapt designs to geotechnical and environmental features. This study shows that soil-structure interaction analysis could be performed with machine learning as advanced decision-making and geotechnical risk mitigation.