High-rate systems refer to structures that undergo rapid changes, exhibiting dynamics that undergo changes in short durations, often less than 100 ms. Examples include hypersonic vehicles, active blast mitigation, and ballistic packages. Developing feedback control systems requires state estimations that can be updated on timescales of less than one millisecond. However, due to the nonlinear and non-stationary dynamics of high-rate systems, they entail high uncertainties, posing challenges for predictive modeling. In this study, we propose a probabilistic machine-learning pipeline for estimating the state of a high-rate system. This approach involves applying probabilistic models and topological data analysis techniques to extract features from the datasets obtained from the Dynamic Reproduction of Projectiles in Ballistic Environments for Advanced Research testbed. We examine the design of probabilistic models for structure state estimation, emphasizing the importance of prediction intervals. We evaluate the best model through several performance metrics, such as mean absolute error, Signal to Noise Ratio, and Time Response Assurance Criterion, while assessing the quality of predictive uncertainty by creating uncertainty calibration curves and calculating the Expected Confidence Error. The incorporation of probabilistic machine learning enables decision-makers to make informed decisions under uncertainty, enhancing the practical utility of the pipeline. The pipeline’s robustness to signal noise and its ability to handle spurious data are presented and discussed.