Carbon fiber exhibits weak interfacial adhesion with the matrix due to its nonpolar nature, which subsequently limits the mechanical strength of the resulting composites. This research examines the chemical surface modification via oxidation, reduction, and silane functionalization to enhance the fiber-matrix adhesion in carbon fiber-reinforced polymer composites. Carbon fiber was treated using HNO3 and H2SO4:HNO3 mixture, followed by reduction and silane functionalization with tetraethyl orthosilicate (TEOS). The presence of hydroxyl (–OH) and silanol (–SiOH) groups on TEOS-grafted fiber was confirmed by Fourier transform infrared spectroscopy. This modified fiber, along with epoxy matrix, was used to fabricate the composite using hand lay-up along with compression molding. The resultant composite was investigated for tensile strength, Young’s modulus, and impact strength. The CFRC-4 composite (sample modified with HNO3 and TEOS) showed highest tensile strength and stiffness, depicting improved interfacial adhesion. Moreover, based on experimental data, the artificial neural network (ANN) model was developed to predict the mechanical performance. At a specific hidden layer size, the optimal performance of stress and strain was precisely predicted by ANN. The experimental validation along with data-driven modeling was combined in this hybrid methodology, providing deeper knowledge of the relation between surface chemistry and composite performance. This approach enables the faster prediction of material characteristics for cutting-edge engineering applications and future research work.