Snow cover varies spatially and inter-annually in the boreal forest landscape due to the canopy influence on snow interception and snow surface energy balance. This is reflected in the soil temperature regime during winter and spring. Differences in the timing of soil warming have been shown to influence the trees� capacity to take up water and assimilate carbon. To understand the effects of forest management or a possible climate change on forest productivity, it is useful to have a tool for predicting the spatial and inter-annual variability of snow depths and soil warming during spring within different stands across the landscape. A SVAT-model (COUP) was used to simulate snow cover and soil temperature during a decade in eight Scots pine (Pinus sylvestris L) stands. The COUP-model proved to be a functional tool for simulating the intra-landscape variation in snow depth and soil temperature. The simulations revealed variability between the stands and between the years, as well as factors contributing to this variability. A more open stand together with low leaf area index resulted in deeper snow layers and, together with higher trees, also an earlier soil warming. The largest spatial variability in the timing of soil warming in spring between sites was found during years with little snow, which is a possible consequence of climate change-related warming in the boreal landscape of northern Sweden.