An increasing number of ontologies demand the interoperability between them in order to gain accurate information. The ontology heterogeneity also makes the interoperability process even more difficult. The existing ontology matching systems are mainly focusing on subject derivatives of the concern domain. Since ontologies are represented as data models in a structured format, in this paper, a new modified model of similarity spreading for ontology mapping is proposed. In this approach, the mapping mainly involves with node clustering based on edge affinity, and then the graph matching is achieved by applying coefficient similarity propagation. This process is carried out by iterative manner, and at the end, the similarity score is calculated for iteration. This model is evaluated in terms of precision, recall, and f-measure parameters, and it is found that it outperforms similar systems.