Wilson’s Disease(WD), a genetic condition, arises from mutations in the ATP7B gene that lead to copper accumulation in the brain and liver. In this study, a multimodal machine learning framework was presented that combines clinical, transcriptomic, and epigenomic biomarkers to enable early and accurate diagnosis. Several models (MLP, DNN, CNN, and LSTM) was implemented for tabular and sequential data, as well as a tabular transformer model and an AutoGluon ensemble. A stratified K-fold approach was utilized for model evaluation and applied SMOTE for class balancing. The performance of models was evaluated using metrics like accuracy, F1, and ROC-AUC. The Tab-Transformer and AutoGluon ensemble method achieved the highest ROC-AUC (0.93 ± 0.02). The SHAP analysis identified body mass index (BMI) and ATP7B gene methylation at specific CpG sites as the most relevant features, while next steps will involve larger external validation studies.