Abstract: The semantic Web technology was introduced into publish/subscribe systems and an intelligent seman-tic matching algorithm (ISMA) was proposed by domain ontology. With double-index hash tables, predicate, variable filter and type tables as the data structure of resource description framework (RDF) graph patterns which stores subscrip-tion conditions, the algorithm adopts meta-statement match counting method to efficiently process atomic subscription factor match only once, and the subscription conditions are an "and" serial sequence matching relationship. The time and space complexities of the algorithm were derived and analyzed by quantitative and qualitative methods. Experi-mental results demonstrate that the algorithm is efficient and scalable, and it is suitable for large-scale publish/subscribe systems.