With the emergence of relatively inexpensive and advanced communication technology, Distributed Database Management Systems (DDBMS) have become an integral part of many computer applications. Efficient query processing is one of the most important issues in distributed database systems. In a distributed environment, it is common that queries extract data from different sites. It is important to limit the amount of data transfer across different sites. Semijoin is a way to reduce the cost of expensive joins between various sites. A key issue in query optimization based on semijoin reduction is to find a good sequence of semijoins that reduce the relations referenced in a given query before the joins are performed. This paper proposes a new approach, based on Genetic Programming (GP), to improve the process of database query in Distributed Database Systems. A longer version of this paper is available.