Technological advancements in the field of artificial intelligence have enabled significant progress in various areas, particularly in optimizing the structural configuration of multi-layer composites. The main objective of this study is to investigate how geometric parameters, such as fiber orientation and number of layers, influence the mechanical properties of these materials. To predict the mechanical properties based on the number and orientation of layers during bending tests, we used a hybrid E-Jaya-ANN optimization technique and compared it with the hybrid Jaya-ANN to evaluate the accuracy of the approach. Additionally, using ABAQUS software, a numerical model has been created based on Hashin’s damage criterion to predict the behavior of composite specimens (CFRP) under bending loads and to collect a number of databases starting with the validation model. Subsequently, we generated a series of numerical results representing various practical scenarios to serve as a basis for training an Improved Artificial Neural Network (IANN). Our ability to obtain a better architecture for the laminated layers was made possible by the influence and variation of these materials’ mechanical properties.