<p>This study employs an integrated Vacuum-Solidification (VS) approach combined with machine learning based predictive modeling to enhance the efficiency of solidifying high water content dredged slurry. The VS method combines a prefabricated horizontal drain (PHD) assisted by vacuum pressure (VP) with a solidification/stabilization (S/S) binder. The study evaluates the effectiveness of VS technique using MgO-activated ground granulated blast furnace slag (GGBS) binders, incorporating flocculants such as APAM and tartaric acid (TA) as additives. Experimental findings revealed that the VS method utilizing the GGBS-MgO binder (denoted as VP-GM) achieved superior strength development, while the inclusion of TA negatively affected early strength due to its hydration-delaying properties. To predict the unconfined compressive strength (UCS), three machine learning models such as Random Forest (RF), RF with Principal Component Analysis (RF-PCA), and RF with Singular Value Decomposition (RF-SVD) were developed and evaluated using multiple performance metrics. The results demonstrated that dimensionality reduction techniques enhanced model generalization, with RF-PCA achieving the most consistent performance across both training and testing phases. Sensitivity analysis identified curing time and GM-APAM content as the most influential factors affecting strength development.</p>