<p>This study develops artificial neural network (ANN) models to predict the axial compressive strength of masonry columns externally confined with carbon or glass fiber-reinforced polymer (CFRP/GFRP) sheets. The work distinguishes itself by providing a comparative evaluation of ANN predictions against both analytical formulations and experimental data, demonstrating improved accuracy, and valuable insights from parametric analysis. A comprehensive database comprising 200 experimental results of CFRP/GFRP-confined masonry columns was compiled from the literature. Multiple ANN architectures with varying numbers of neurons in a single hidden layer were tested to optimize performance. The ANN16 model, with 16 neurons in one hidden layer, achieved the best predictive accuracy. The model’s performance was validated against both experimental data and traditional analytical approaches using statistical metrics such as mean squared error and coefficient of determination (<i>R</i>²). The proposed ANN model demonstrated strong agreement with experimental results (<i>R</i>² = 0.88) and outperformed existing analytical models (<i>R</i>² = 0.83). In addition, a detailed parametric analysis was conducted to examine the influence of column geometry, masonry material strength, and CFRP/GFRP characteristics. These contributions establish ANN models as an effective and practical tool for predicting the axial compressive strength of CFRP/GFRP-confined masonry columns.</p>