Purpose <p>Prediction of natural frequencies of thin plates with cutouts is essential for effective vibration control and structural optimization in aerospace and automotive applications. This study investigates the modal behaviour of AA2024-T3 aluminium plates with a single central cutout and proposes a rapid and reliable framework for modal analysis by integrating finite element analysis (FEA) with machine learning (ML) techniques.</p> Methods <p>Plates with varying geometric parameters and cutout shapes (circular, square, and oblong) were analysed under six boundary conditions using ANSYS to generate a comprehensive dataset of natural frequencies. Several ML models, including Linear Regression, Polynomial Regression, Random Forest, Gradient Boosting, and Artificial Neural Network (ANN) were trained and evaluated to predict modal frequencies.</p> Results <p>Fully clamped plates exhibited the highest natural frequencies, whereas plates with free edges showed the lowest. Among the different cutout geometries, circular cutouts retained structural stiffness more effectively than square or oblong ones. Among the tested ML models, the ANN model outperformed all others, demonstrating excellent predictive accuracy with an R² value of 0.9998, indicating a very strong correlation with the finite element analysis results.</p> Conclusion <p>The proposed FEM–ML framework offers a fast, accurate and computationally efficient approach for predicting vibrations in plates with central openings, facilitating rapid structural optimization and informed design decisions in practical engineering applications.</p>