An accurate description of material properties and boundary conditions is the prerequisite to achieve reliable results of a hygrothermal performance simulation. However, the uncertainties widely exist in those input variables, e.g., due to the inhomogeneous nature of materials and discrepancies in material manufacturing and measurement processes, the material properties are subjected to considerable variation. Therefore, the simulation result may not be only a single value, but in a range of possibilities. In this article, a stochastic approach is developed and implemented by using uncertainty analysis which assesses the uncertainties of a model's input variables on that of the output variables, and sensitivity analysis, which identifies the most influential input variables. The hygrothermal performance of one typical wall assembly in North America is studied by applying this approach, considering the uncertainties of material properties of each individual layer and the uncertainties of the interior and exterior boundary coefficients. The influence of wall orientation is also presented. The results show that the effect of a single input variable on an output variable is not constant but varies with time. The key variables that affect the results are also different over time.