Chloride ingress into reinforced concrete (RC) structures exposed to aggressive environments is a major factor contributing to steel corrosion and long-term durability loss. Analytical models based on Fick’s law are widely used but rely on strong assumptions, limiting their applicability. Numerical methods such as Finite Element Method (FEM) or Finite Volume Method (VFM) address complex conditions more effectively but are computationally intensive and require detailed inputs. Data-driven approaches using deep learning offer flexibility but often lack physical interpretability and depend on large datasets. To overcome these limitations, this study proposes a Physics-Informed Neural Network (PINN) framework that incorporates the governing diffusion equations into the learning process. The model does not require extensive experimental data but allows integration when available. By embedding physical laws directly into training, the proposed approach ensures physical consistency, enhances generalization, and reduces data dependence. This work offers a reliable and efficient alternative for modeling chloride diffusion in RC structures, particularly in scenarios where data are scarce and physical fidelity is essential.