The rupture of aortic aneurysms is a life-threatening and most fatal condition and predicting them is a major concern in clinical practice. The traditional risk assessment protocols are mainly based on the diameter of an aneurysm, which does not capture the biomechanical and clinical variance in patients adequately. This paper suggests a machine learning-based architecture of rupture risk forecasting in aortic aneurysms through combination of imaging biomarkers using computed tomography angiography with a detailed clinical history. The data set was analysed on a retrospective basis of 620 patients and includes morphological and biomechanical imaging features in addition to demographic as well as clinical variables. Four monitored learning schemes; Logistic Regression, Support Order machine, Random forest, and Extreme Gradient Boosting (XGBoost) were composed and assessed. Experimental findings indicated that the ensemble-based models performed better than the linear models where XGBoost had the best performance of not only 89.4% accuracy, sensitivity of 0.86, specificity of 0.91, but also an area under the receiver operating characteristic curve (AUC) of 0.92. The analysis of the importance of features the most prominent predictors of rupture were found to be peak wall stress, aneurysm diameter, intraluminal thrombus volume, and growth rate. The framework presented in the research demonstrated a significant improvement in predictive accuracy over traditional methods that rely on the diameter-based and statistical methods. The results reveal the clinical possibilities of multimodal risk stratification based on machine learning to aid personalised decisions and outcomes in the treatment of aortic aneurysms