Abstract: In the research of ontology mapping，ontology block mapping is the most popular of large-scale ontology mapping task.However，current partition methods have the following drawbacks：the size of block is not suitable for mapping，partition quality relies on the prior parameters.Aiming at solving those problems，this paper considers both cohesion and coupled degrees which guide the clustering process in order to improve the partition quality.Besides，this paper re-designs ontology block mapping method based on cover clustering with parameters unrelated.Experiment result shows that the method developed is fairly efficient for solving large-scale ontology mapping tasks.