Accurate reconstruction of epidemic dynamics is challenging when reported infection data are incomplete or affected by significant under-reporting. Excess mortality indicators provide an alternative source of information that can be used to infer epidemic trajectories. In this study, we propose a regularized inverse calibration framework for a SEIRD epidemi-ological model using excess mortality data. The calibration problem is formulated as an inverse problem and stabilized through a Gaussian functional regularization that constrains the admissible epidemic trajectories. This approach reduces sensitivity to noise in mortality observations and prevents oscillatory solutions typically associated with ill-posed param-eter estimation. The model is numerically integrated using a fourth-order Runge–Kutta scheme and calibrated against mortality data from Catalonia. Cross-context validation is further performed using mortality data from Ecuador to assess the structural robust-ness of the approach. The results show that the regularized calibration produces smooth and epidemiologically consistent epidemic trajectories while maintaining agreement with observed mortality patterns. The proposed framework provides a robust methodology for reconstructing epidemic dynamics from mortality indicators and may contribute to improved epidemiological surveillance in situations where case reporting is limited or unreliable.