In the present study wildland fire occurrence pattern is studied in two well-documented test-sites in Switzerland and Greece. The emphasis of this study is to assess the explanatory performance of landscape variables and to investigate the potential of integrating spatial associations for explaining long-term wildland fire occurrence patterns. In general, predictive regression models make the assumption that the observations are independent of each other. However, there is evidence that fire ignition locations can be (auto-) correlated over space and time. Consequently, spatial measures and dependencies can be introduced to minimize prediction residuals and thus to improve the explanatory performance of the independent variables for explaining wildland fire occurrence pattern. The hypothesis is that regions of similar or equal environmental conditions may show different fire occurrence pattern because of different neighborhood effects.