Despite the benefits of wood as a sustainable raw material used in the construction field, the changes of moisture content (MC) inside wood can be a risk for the safety, serviceability, and durability of timber components exposed to variable climates. This work proposes a multi-phase finite element model based on X-ray computed tomography (CT), able to simulate the moisture transport in wood in the hygroscopic and over-hygroscopic range. The general form of a multi-phase equation is shown in Eq. 1 where c, D, and t are the water concentration, the diffusion tensor, and the time, respectively, and dot c is a conversion rate. The subscript w refers to one water phase (e.g., free water in pores) while b and v refer to the other water phases (e.g., bound water in wood cell walls and water vapour in pores): (∂c_w)/∂t = ∇∙(D_w ∇c_w) + c ̇_wb + c ̇_wv (1) Compared to recent multi-phase models, the conversion rates are defined on the basis of the CT measurements. The model is successfully validated for the case of small cylindrical pine specimens initially saturated and then subjected to drying until a relative humidity RH=33%, wetting until RH=94%, and drying until RH=64% at a constant temperature of 23 °C. Then, in the hygroscopic range, the moisture transport analysis is sequentially coupled with a mechanical analysis previously used where the shrinkage and swelling parameters are now defined based on CT that allows to analyze the cell wall structure and its changes under different MCs (e.g., cell wall thickness and deformations) in various wood locations (e.g., heart-, sap-, early-, late-wood). The proposed model and the whole CT-based methodology are promising to assist the largely used sensor-based monitoring techniques for wooden components.