Acetylation is a chemical treatment method commonly used to improve the hygroscopic properties of wood. Although acetylation has been industrially used for decades, its effects on the different hierarchical structures of wood are still poorly understood. In the laboratory, acetylation is generally measured gravimetrically. Weighing a sample before and after the modification procedure provides an indirect measure of the degree of acetylation within the entire sample but does not provide detailed information on the different structural regions of wood. Here, we determined acetylation of wood surfaces using hyperspectral near-infrared image regression. Our results show significant differences in the acetylation of earlywood and latewood, which suggests different durations for complete acetylation of earlywood and latewood cells. We have also illustrated our findings on the wood cell level based on the chemical differences in earlywood and latewood cell walls using cluster analysis of Raman images. These findings are an important step in understanding how chemical treatment affects the different hierarchical structures of wood on different spatial scales.