This study proposes a thermodynamic characterization method for reservoir water sensitivity based on macro-lens infrared thermal imaging, aiming to address the lack of an efficient and fully quantitative technique for evaluating the microscopic heterogeneity of the mixed layer ratio of illite/smectite (I/S). The I/S mixed layer is the dominant contributor to reservoir water sensitivity in extensively developed deep tight reservoirs in China and a core geological factor leading to generally low productivity of such reservoirs. A quantitative relationship between the I/S mixed layer ratio and apparent temperature is established in this study, which confirms that a lower apparent temperature corresponds to a higher content of smectite layers and stronger lattice expansion potential. Using this method, we reveal for the first time the ubiquitous thermal gradient characteristic of “hot edges and cold center” in I/S aggregates, and verify from a thermodynamic perspective that illitization is a centripetal metasomatic process controlled by fluid-mineral interfacial diffusion. On this basis, a three-level thermodynamic evaluation criterion for the water sensitivity potential of the I/S mixed layer is proposed, which enables rapid identification and quantitative classification of microscopic reservoir damage risks. The findings break through the technical bottleneck that traditional methods cannot efficiently and quantitatively characterize the planar heterogeneity of the I/S mixed layer ratio, and provide a fast, non-destructive and high-precision new method for reservoir water sensitivity evaluation. It will also provide a brand-new theoretical basis for sweet spot optimization and reservoir protection in tight oil and gas reservoirs.