In this paper, a new model for chloride diffusion in cracked concrete is presented. The model was developed within a standard continuum damage mechanics framework but newly considers both an irreversible damage variable and a strain field, gaps identified through a literature review, offering a more comprehensive description of the diffusion-like processes in concrete with small and moderate cracks than current models. The model utilizes a realistic dependence for the diffusion coefficient on the damage state and employs a bi-logistic mapping function or an S-curve mapping function to effectively describe the non-linear effects of cracks on diffusion in concrete. The dependence on the strain field introduced by this model not only allows examination of diffusion in opening cracks, it also enables the process of crack closing to be simulated, whatever the irreversible damage variable may be. This paper also proposes a novel calibration procedure for experimental testing involving the use of the Brazilian splitting test on samples enclosed in a steel ring. This innovative approach simplifies the process of creating a portfolio of cracks with varying widths. In the calibration tests described here, samples were subjected to natural diffusion while chloride content was monitored in inlet and outlet chambers attached to the samples, eliminating the need for laborious chloride profiling. Our in-depth examination of chloride diffusion in cracked concrete is applicable to analyses of corrosion in steel-reinforced concrete structures worldwide, with the new calibration procedure offering a straightforward method for populating the model with realistic material parameters, thus providing a valuable tool for accurately predicting chloride penetration in real-world scenarios.