To process queries against a global schema in a distributed multidatabase system, query transformation which maps the query against the global schema to the local schemas is necessary. In this paper, a localized approach is presented which optimizes distributed query processing. This approach transforms a query to subqueries. Each subquery references data contained at a local site and therefore can be locally processed. It reduces data transmission time and query response time due to local processing and parallel processing. It can therefore be applied in distributed database systems implemented in either long-haul or local networks. Moreover, semantic information can be incorporated with this approach to process distributed queries at a single site. Data fragmentation and schema and data conflicts exist in the multidatabase environment, which are treated in a uniform framework in this approach.