Due to the natural growth process of trees, wooden boards show a high amount of random fluctuation in their mechanical properties. Currently, in engineering practice, a homogeneous material behaviour is assumed within wooden boards and uncertainties in loading and load-bearing capacity are considered through design procedures based on partial safety factors. These partial safety factors are not directly linked to the individual mechanical behaviour of the considered structural system/element. Thus, a stochastic framework opening up the possibility to establish such links, by combining structural analysis tools with probabilistic descriptions of wooden boards, could be an important step in timber design, also to improve the competitiveness of wood against other building materials.In this study, a framework able to consider the impact of random stiffness fluctuations of wooden boards on the performance of glued laminated timber (GLT) is presented. A 2D finite element model is employed to depict the interactions between individual laminations. It is tightly integrated into a commercial robustness analysis and optimisation software package.The mechanical properties of each board are obtained on basis of measurements collected during the grading process and are condensed into so-called stiffness and strength profiles. Based thereon, a probabilistic material model is developed for the random generation of an arbitrary number of stiffness and strength profiles. On basis of the probabilistic material model and the mechanical FE model, the sensitivity to changes in the design can be explored. Subsequently, an adaptive response surface approach leads to the optimal design parameters of the user-defined structure.An exemplary application scenario shows the workflow for determining the optimum hole dimensions and location in a GLT beam. The material model indicates large scatter in the mechanical properties, leading to conservative results. Nonetheless, and although only bending failure of GLT beams and no finger joints are taken into account, the example shows the strengths of the employed workflow from sensitivity analysis to best designs.