AbstractThis paper compared several mathematical models predicting chloride penetration in concrete by employing a time-dependent diffusion coefficient for a number of analytical solutions of Fick’s second law. We propose calibrated models that consider a time-dependent build-up of the surface chloride concentration Cs. Predicted values and measured values of chloride profiles from experimental results were compared, which showed that the calibrated models considering a time-dependent surface chloride concentration could significantly improve the correlation compared with the conventional models employing a constant surface chloride concentration. Moreover, we investigated several factors (i.e., skin effect, aging factor m, exposure conditions, curing conditions, water/binder ratio, and fly ash content) that may affect chloride penetration in concrete by adopting the calibrated models. The results showed conclusively that these factors also significantly affected the accuracy of the chloride penetration obtained by these calibrated models. Finally, the results predicted by the calibrated Tang–Gulikers model were proved to be reasonably accurate in concrete with a small water/binder ratio compared with the experimental data. It is concluded that the calibrated Tang–Gulikers model can be applied for predicting chloride penetration and solving relevant service-life problems resulting from chloride ingress in concrete.