Cellulose forms crystalline fibrils, via biosynthesis, that can be just a few nanometers wide. The crystallinity is a structural factor related to material performance. Recently, many routes to isolate these fibrils as nanocellulose have been developed, and there exist various types of nanocellulose with different crystallinities. Quantitative assessment of the crystallinity of nanocellulose is thus essential to advance knowledge in the high performance and functionality of such materials. Solid-state 13C cross-polarization/magic-angle spinning (CP/MAS) nuclear magnetic resonance (NMR) spectroscopy is a strong tool to investigate the structural features and dynamics of solid cellulose. The crystallinity is often evaluated by using the NMR signal ratio of the C4 crystalline and noncrystalline regions as a crystallinity index (CI) value. To calculate the CI value, it is necessary to examine the dependence of the contact time (CT) for CP on the signal intensity and set the optimum CT at a maximum of the signal intensity. However, the dependence has not been investigated for evaluation of the CI value of various cellulose samples with different crystal sizes. Here, we optimized the CT for evaluation of the CI value of cellulose with different crystal sizes. The error in the CI at the optimized CT was ~ 3%. At the optimized CT, the structural change after surface modification by TEMPO-oxidation was also analyzed from the NMR spectra of the C6 region. The relationship between the CI value and the degree of oxidation shows that it is possible to evaluate the degree of oxidation from the NMR spectra irrespective of the crystallinity of cellulose. Furthermore, the C4-based CI value was linearly correlated with the C6-based trans-gauche (tg) ratio, which is approximated by a function, CI = 0.9 tg ratio.