Even if many studies in the literature confirm that the thermal properties of several building materials are highly affected by the material’s temperature and moisture content, it is common practice in building performance simulation to adopt constant properties, such as thermal conductivity. Those values are usually estimated according to the technical standard EN 10456:2007. This modelling choice can introduce inaccuracies in the calculation of the heat and mass transfer across the opaque building envelope, especially for some types of buildings such as the timber ones. With the goal of assessing and quantifying the extent of these inaccuracies in timber components, this research combined experimental and numerical analyses. First, dynamic thermal conductivity functions dependent on temperature and moisture content were determined for some typical timber materials (cross-laminated timber, woodchips, and wood-fibre panels). Then, annual hourly simulations were run for a set of different timber components in a Mediterranean and an Alpine Italian climate, considering four different approaches of increasing complexity and level of detail for the determination of the thermal conductivity values. Results reveal that the current standard approach based on the constant thermal conductivity determined in agreement with the EN 10456:2007 conditions can lead to significant inaccuracies compared to more detailed alternatives, whose values can be affected by the climate, the component’s position, and orientation, as well as its solar absorptance. The heat flux across a horizontal roof placed in a warm and humid Mediterranean climate can increase up to a +9 % when considering a dynamic thermal conductivity instead of the constant approach.