Aerospace components must be designed to be damage tolerant, which typically involves applying large knock-down factors to account for a worst-case undetected damage between inspections. High-fidelity Finite Element Method (FEM) models can be used not only to mitigate these uncertainties, and thus knock-down factors, but also to gain insight into the progressive damage behaviour [1]. However, the high computational time and associated costs of these high-fidelity FEM models limit their applicability in design and optimisation. To address this limitation, Machine Learning (ML) can be employed to fast obtain aggregated damage quantities for a wide range of model parameters [2]. This work presents the development and validation of a high-fidelity 3D impact FEM model against experimental testing. In a subsequent step, the model was simplified by assuming 2D axi-symmetry, reducing the computational time and enabling the generation of a database comprising 128 datapoints, along with a test dataset of 16 datapoints. A ML method was then developed and trained on this dataset. This model employs an ensemble of Multi Layer Perceptrons. This ensemble both improves the prediction compared to one network only, but can also be used to indicate the uncertainty in the predictions. The developed methodology enables fast prediction of impact damage behaviour, facilitating its use in design, multi-disciplinary design optimisation, and digital twin development, among other potential applications. Future work will involve expanding the range of input parameters, further refining the high-fidelity FEM model, and investigating alternative Machine Learning (ML) methods to enhance the predictive capabilities of the developed methodology. [1] N. van Hoorn, S.R. Turteltaub, C. Kassapoglou and W.M. van den Brink. Numerical prediction of impact damage in thick fabric composite laminates. Composite Structures, Vol 353, pp 118726, 2025. https://doi.org/10.1016/j.compstruct.2024.118726 [2] H. Ji, Y. Zhang, X. Wang, L. Qin, Z. Yue, B. Li, Z. Li, H. Meng, P. Wang, R. Zhang and T.J. Lu. A few-shot deep learning framework for predicting high-velocity impact response of ultra-high molecular weight polyethylene fiber-reinforced composites. Aerospace Science and Technology, Vol 163, pp 110152, 2025. https://doi.org/10.1016/j.ast.2025.110152