<p>Fire resistance is a key aspect that determines the structural integrity of building columns during exposure to high temperatures. Traditional experimental and analytical approaches to the fire resistance determination of RC and CFST columns are often too time-consuming and costly, with reduced predictive capability under various fire scenarios. This paper aims to develop ANN models with accuracy and efficiency for the fire resistance prediction of RC and CFST columns based on geometric and material properties. Two comprehensive datasets were compiled from the literature, involving 270 RC and 223 CFST column specimens that have undergone fire resistance tests. Eight input parameters have been considered for RC columns, and twelve parameters for CFST columns, which include geometric dimensions, material strengths, characteristics of the applied load, and heating rate. The ANN models were trained and assessed for their performance by correlation coefficient (R), root mean squared error (RMSE), and mean squared error (MSE). Optimum ANN model demonstrated R values of 0.92294 and 0.99431 for RC and CFST columns, respectively, thus showing better results compared to existing empirical models. Parametric analyses further highlighted the influence of critical parameters such as heating rate, load eccentricity, and concrete strength on fire resistance. The developed models provide a practical, data-driven approach for rapid and reliable prediction of column fire performance, offering substantial value for structural design, fire safety assessment, and performance-based engineering applications.</p>