MRI of fixed post-mortem tissue can be used to relate MR signals to histopathology. However, MR relaxation parameters are altered by the process of fixation. We propose a “kinetic tensor” model that simulates the influx of fixative into the brain using diffusion tensor data, and demonstrate a correction by fitting a fixative concentration map to quantitative T2 measurements. Compared to correction methods based on distance-to-surface or isotropic diffusion, our approach captures an individual brain’s morphology and microstructure, both of which influence fixation. T2 maps corrected with the kinetic tensor model are more homogeneous than those with alternate corrections.