The solid-state NMR assignments of the C-13 resonances of bacterial cellulose IR were reinvestigated by INADEQUATE experiments on uniformly C-13-enriched samples from Acetobacter xylinum. Additionally, we determined the principal chemical shift tensor components of each C-13 labeled site from a 2D iso-aniso RAI (recoupling of anisotropy information) spectrum acquired at magic angle spinning speed of 10 kHz. On the basis of these NMR data, the crystal structure of cellulose IR was refined using the C-13 chemical shifts for target functions. Starting off with coordinates derived from neutron scattering, our molecular dynamics simulations yielded four ensembles of 200 structures, two ensembles for hydrogen bond scheme A and B and two ensembles for different chemical shift assignments I and II, giving 800 structures in total. These were subsequently geometry-optimized with the given isotropic chemical shift constraints applying crystallographic boundary conditions, to identify a structure for every ensemble that fit best to the experimental NMR data. The resulting four model structures were then assessed by simulating the chemical shift tensors (using the bond polarization theory) and comparing these values with the experimental chemical shift anisotropy information (obtained by RAI). The earlier neutron diffraction study had reported two possible occupation schemes for the hydrogen-bonded hydroxyl-groups (A, B) which connect the cellulose chains. From these two possibilities, our NMR results single out pattern A as the most probable structure. In this work, the first time crystallographic boundary conditions were applied for C-13 chemical shift structure refinement for molecular dynamics simulations and Newton-Raphson geometry optimization.