Wood hardness is an important property for flooring, furniture products and structural utilization. Currently wood hardness can only be measured using destructive testing. As there is no suitable method for predicting wood hardness from standing trees, developing a non-destructive technique to predict wood hardness from plantations trees would provide significant benefits for evaluating optimal silvicultural treatments, and for selecting trees for tree improvement programs. It is proposed initillay that a possible non-destructive test may be developed using „crystallinity“ characteristics (degree of crystallinity (DC), crystalite width (CW) and length (CL) and microfibril angle (MFA)) determined using X-Ray diffraction. The aim of this study was to test the feasibility of using crystallinity as a non-destructive indicator of wood hardness. Experimentally, nine trees were taken representing small, medium and large diameter from five-year old fast growing teak plantation forest in Ciampea, Bogor, Indonesia. Wood hardness and crystallinity were determined following a standard testing regime. The results showed that MFA was negatively related to all hardness directions. Crstallite width only influenced side hardness with negative correlation, while degree of crystallinity only affected the end-grain hardness with positive correlation. Different vertical positions within tree only influenced the end-grain hardness value, while different radial positions had no effect on wood hardness. The optimum sampling height was found to be 80 cm, 130 cm, and 100 cm for radial, tangential and end-grain hardness, respectively. The relationship was based on the height that provided the highest correlation with the whole tree hardness values. Proposed models involving density and „crystallinity“ for predicting wood hardness are presented. While the R2from the proposed models was mostly less than 0.50, however, the approach described provides a new way to predict wood hardness from 10 mm increment cores at standing trees.