A machine learning-based calibration strategy for the ITER Tokamak Systems Monitor (TSM) is presented, focusing on the optimization of dynamic finite-element method (FEM) models essential for monitoring and assessing the structural health of the tokamak. Traditional finite-element model updating techniques are unsuitable for the ITER environment due to constraints like limited sensor placement and the inability to perform controlled excitations. To overcome these challenges, an automated calibration procedure is proposed, leveraging Bayesian optimization with Gaussian Process regression. The parameter space is efficiently explored by sequentially sampling model configurations, minimizing the discrepancy between the simulated and reference modal characteristics. This strategy was applied to calibrate a simplified 40°FEM model of the tokamak using synthetic data from a comprehensive 360°FEM model acting as a proxy for the real machine. A significant alignment between the models’ natural frequencies was achieved using a minimal number of simulations. The approach reduces the need for manual calibration and is computationally efficient, making it suitable for complex and costly numerical models. While this study focuses on dynamic FEM models, the proposed method establishes a foundation for calibrating other numerical models within the TSM. This preliminary work demonstrates the potential of Bayesian optimization for model calibration in tokamak systems, offering a scalable and effective solution under practical constraints.