Abstract. Adaptive Partitioning Algorithms (APA) divide the feasible region into non-overlapping partitions (regions) in order to direct the search to the promising region(s) that are expected to contain the global optimum. APA usually collect data from pre-determined locations in each partition and use evaluation measures that are based on assumptions or function approximations. The proposed Fuzzy Adaptive Partitioning Algorithm (FAPA) is a novel approach that aims at locating the global optimum of multi-modal functions without using any assumptions or approximations. FAPA introduces two new features: it selects the locations of data randomly in each partition and it utilizes a fuzzy measure in assessing regions.