The need and importance of analysing and comparing maps generated by modelling tools (e.g. habitat suitability models) has been growing among researches to detect temporal/spatial changes, compare different scenarios, to analyse model sensitivity or to calibrate/validate models. To analyse model performance, usually a cell-by-cell comparison including an error matrix is used, which allows an application of Kappa-statistics to derive an overall agreement between observed and predicted maps. However, cell-by-cell agreement between two maps only contains information about the specific cells. Thus, it fails to differentiate between a near-miss and a far-miss, spatially. To overcome this deficit, the Kappa-fuzzy-method is going beyond cell-by-cell comparisons and gives partial credit to cells found in the neighbourhood. This means, when matching attributes are found at shorter distances the agreement is higher. In this study, the Kappa-fuzzy-method compares maps with observed spawning areas with maps of simulated spawning areas for European grayling (Thymallus thymallus) and maps of simulated juvenile habitats for brown trout (Salmo trutta) which have been simulated using different physical-biota relationships (e.g. fuzzy-method, preference functions). Both, fuzziness in location (neighbouring cells) and fuzziness in category (low, medium or high habitat suitability) are investigated. The results show an increase in performance, when fuzziness in location and category are considered. Hence, the Fuzzy-Kappa-Method allows to consider the degree of uncertainty involved in habitat modelling. However, care has to be taken in defining the degree of fuzziness because it represents a strong feature to artificially increase agreements beyond physical and biological means.