This study applies multiple machine learning techniques, including Gaussian Naive Bayes (GNB), Support Vector Machines (SVM), Random Forest (RF), and Linear Regression (LR), to predict the intensity and risk level of seismic hazards such as earthquakes and tsunamis. The goal is to provide accurate predictions of magnitude and risk levels to mitigate potential losses. The framework integrates various algorithms tested on a comprehensive dataset of recorded seismic phenomena, accounting for factors such as latitude, longitude, disaster type, focal depth, and water height. Models are trained to predict future magnitudes and risk categories, which are crucial for estimating potential damages and planning appropriate responses. The performance assessment considers metrics including accuracy, precision, recall, and F1-score for the classification models (GNB, SVM, RF), while the Linear Regression model provides specific magnitude predictions. The system employs an ensemble approach, combining probabilities from multiple models to determine risk levels and alert statuses. This multi-model strategy aims to provide more robust and reliable predictions by leveraging the strengths of different algorithms. The framework also includes data preprocessing steps such as handling missing values, encoding categorical variables, and scaling features to improve model performance. This comprehensive approach allows for a nuanced understanding of seismic risk, categorizing it into different levels (Low, Moderate, High, Extreme) based on predicted magnitudes and probabilities, thereby offering valuable insights for disaster preparedness and response planning.