<p>The Present work aims to develop and validate Artificial Neural Network (ANN)-model to determine the deformation behaviour of hemp fibre-reinforced with epoxy composites prepared through hand lay-up method (60:40 fibre to matrix ratio). To produce a credible set of data which would represent standard and load bearing as well as defect sensitive responses, experimental tensile and edge notch tensile (ENT) tests were performed according to ASTM standards. ANN with structured multi-layered perceptron, trained on mechanical and geometrical input parameters, which indicates effective model convergence and high predictive accuracy, with an R² value approaching 0.95 and low overfitting. The model was able to reproduce the nonlinear deformation patterns experimentally and it therefore allowed accurate simulation behaviour of composite under distinct loading and dimensional conditions. The modelling framework was also supported by complementary SEM–EDS analysis to ensure that there was the fibre-matrix combination, and consistent elemental distribution throughout the interface, which ensured the integrity of the dataset used in training. Combining experimental validation with the data-driven prediction, the study shows that ANN-based methods can be used to lower the utilization of the massive physical prototyping, increase the level of computational efficiency, and assist in sustainable and performance-based composite design. The results indicate how machine learning can be of great importance in the evolution of predictive modelling and intelligent engineering of natural fibre-based polymer composites.</p>