Confining structurally deficient concrete columns with externally bonded fiber-reinforced polymer (FRP) has been widely accepted as an effective technology for strengthening the ductility and strength of deficient concrete columns. However, prediction models for damaged and afterward repaired concrete based on soft computing methods are not available for the planning and maintenance of concrete structures. Therefore, this paper adopted two soft computing methods – artificial neural network (ANN) and Gaussian process regression (GPR) – to analyze observations obtained from 103 datasets of concentrically loaded FRP-confined predamaged concrete. The models only consider statistically significant variables with the ultimate strength of FRP-confined predamaged concrete. The statistically significant variables based on the multivariate regression analysis are corner radius ratio, FRP thickness, concrete strength, and damage degree. The coefficient of determination of the developed models is greater than 98% and there is a relatively low error between the measured and predicted values. The results of the current study highlight the merit of using soft computing methods in concrete technology given their extraordinary ability to comprehend multidimensional phenomena of concrete structures with ease and high predictivity over the existing empirical models.