A moisture content effect on fracture characterisation of Pinus pinaster under mode I is addressed. The double cantilever beam test is selected for mode I loading, based on specimens scaled down to the growth ring level. Specimens are stabilised using aqueous solutions at several equilibrium moisture contents () ranging from 0 % to about 13 %. The strain energy release rate () is evaluated by applying the compliance-based beam method, from which the Resistance-curve is determined directly from load-displacement (P-delta) data without crack length measurement. The crack tip opening displacement in mode I () is determined by post-processing the displacements at the initial crack tip determined from digital image correlation. and are then combined for the direct identification of the cohesive law, , which is assumed in cohesive zone modelling. Experimentally, crack propagation consistently occurs in the earlywood layer. The increase of does not show any influence on the initiation stage of the fracture process zone. However, a statistical correlation exists for the critical () and at maximum load () values of with regard to . Consistently, the area under the identified cohesive curve increases with , although high scatter and low correlation between maximum cohesive strength () and are observed. The methodology is also validated using finite element analysis including cohesive elements and taking into account the growth rings heterogeneity. The numerical results show that the identification of the cohesive law is insensitive to the variability of the growth ring structure observed experimentally.