This work presents a numerical model to analyse the hygro-thermal behaviour of wooden bridge members. A multi-Fickian hygro-thermal model, previously implemented by some of the authors, is extended by including the dependency of wood sorption on temperature above and below zero degrees Celsius to predict moisture, temperature and relative humidity in wood under Northern European climates. The performance of the model in the presence of protective paints is particularly investigated. The finite element analysis based on the proposed model simulates the hygro-thermal behaviour of a glue-laminated beam of Älsvbacka Bridge located in Skellefteå (North of Sweden). The beam, coated by paints and claddings, was monitored by using wireless sensors in a previous research. Comparisons with the available measurements reveal that the numerical model is able to predict the moisture content in locations sheltered from rain and sun with moisture levels below the fibre saturation point. A study of the influence of different protective paints shows that the maximum and minimum moisture content at various depths along horizontal paths of the beam cross section, as well as the moisture gradients in different seasonal periods, are strongly affected by the type of paint. The proposed numerical approach is a promising tool to facilitate sensor-based monitoring techniques and to optimize the choice of protective paints for improved performance of timber bridges and other wooden structures under variable climates.