Abstract: A fuzzy membership function was defined for each direction to predict the membership degree that an object pertains to a certain direction. Nine fuzzy membership functions were defined to describe the direction relations between fuzzy objects and crisp objects with corresponding fuzzy sets. Direction relations were described by a 3×3 fuzzy matrix, which was computed by an aggregation operator defined on the nine fuzzy sets. The fuzzy matrices and crisp matrices of direction relations between fuzzy objects and crisp objects were computed respectively, and comparison of fuzzy matrices with crisp ones reveals that the proposed fuzzy approach is more effective than existing crisp method.