We assessed climate change impacts on wildfires using fire modeling approach for three 30-years periods in the Mediterranean island of Sardinia (Italy). Climate projections based on the A1B IPCC emissions scenario have been carried out using the CMCC-CLM regional climate model (Rockel et al. 2008, spatial and temporal resolution of 14 km and 6 hours respectively) to derive the input data for fire spread modeling. The three studied periods were: baseline (1981-2010) and two future periods (2011-2040 and 2041-2070). To characterize the impacts of these projected changes in climate and fuel moisture conditions on fire behavior and exposure we applied a fire spread model based on the Minimum Travel Time algorithm (MTT, Finney 2002) called RANDIG. The fire simulations were run at 250 m of resolution, considering a set of 50000 fire ignitions randomly sampled from the historical database provided by JRC. To evaluate the fire exposure variations only due to climate changes, we supposed no changes in land cover and in ignition patterns for the fire simulations. We then derived and analyzed fire exposure profiles (burn probability, conditional flame length, fire size) for the whole island. Results evidenced interesting spatial and temporal variations among the three studied periods, for both climate conditions and wildfire exposure, in the study area. In particular, the main findings of the work for future conditions in Sardinia were the potential lengthening of the fire season, the likely reduction of the days characterized by strong wind intensity, and the limited variations in terms of fire exposure factors. This methodology allows developing guidelines for fire management and prevention policies considering the future climate change projections in Mediterranean areas, and could be easily extended to other areas.