The rise in computational power has encouraged researchers to use concepts like machine learning in the field of computational mechanics. In this work, artificial neural networks (ANNs) are applied in the field of stress recovery. Stress recovery is a widely studied topic since the day finite element method (FEM) began. Machine learning models are trained using stresses from finer meshes. The trained model is then used to calculate stresses at any point over the domain. These stresses are compared with stresses calculated in a coarse mesh, from the popular stress recovery technique, the superconvergent patch recovery (SPR) technique. Errors are calculated by considering stresses from a very fine mesh as reference stress. Three separate models are created for the three components of stress in a plane stress problem. Using the developed ANN models, the Cook’s skew beam problem is solved.