The service life of buildings and components is extremely difficult to model and predict, mainly due to the uncertainty and variability associated with the degradation phenomena. These uncertainties are usually related with the intrinsic randomness of natural phenomena, such as the environmental degradation agents. In order to deal with these uncertainties, in this study, a stochastic approach is used, to model the service life of timber claddings. For this purpose, a logistic regression analysis is used to evaluate the probability of timber claddings to reach the end of their service life over time. Moreover, the impact of the environmental exposure conditions (e.g. facades orientation, distance from the ocean, exposure to damp and exposure to wind/rain action), which influence the degradation and service life of timber claddings, is also evaluated. The proposed models are applied to 150 timber claddings, located in Valdivia (South Chile), which have been analysed during an extensive fieldwork survey. The results obtained in this study aid the definition of maintenance plans, in a technically informed manner, in order to promote the conservation of these claddings, which are an important symbol of the local identity of this region, depicting a complexity of cultural values, inherited by Spanish and German colonisations.