Using fire detection data from remote sensing satellite imagery, this study examines wildfire distribution across Italian municipalities during the summer of 2023. Predicting wildfire occurrences at the local scale remains complex due to their broad spatial variability and the nonlinear nature of fire behavior. To address these challenges, we propose modeling satellite-based wildfire data using spatial models while addressing the change of support issue through methods that enable the integration of both point- and area-level analyses. This strategy allows for the use of the Integrated Nested Laplace Approximation (INLA) framework for wildfire estimation. By accounting for zero-inflated wildfire counts, the analysis captures local variability and highlights the influence of environmental and socio-economic factors. Model validation relies on leave-group-out cross-validation to enhance estimate reliability and ensure robustness of results. Ultimately, the study underscores the challenges of integrating data at different scales for local-level analysis, advocating for more effective use of multi-source data in future wildfire monitoring efforts.