The Genebank Project of the Ministry of Agriculture, Forestry and Fisheries, Japan established in 1985, has developed an extensive network for collection, classification, preservation, evaluation, utilization and documentation of genetic resources preserved in Japan. Among these activities, evaluation and documentation are of particular importance for breeding programs. A database system has been established with the purpose for promoting efficient utilization of the genetic resources contained in the Genebank. This paper presents an outline of the database on fruit trees. The characters of fruit trees are divided into three categories: primary or morphological: secondary or physiological and ecology: and third or agronomic and qualitative characters. The number of primary characters ranges between 11 to 53, depending on kinds of fruits. The evaluation data on various characters are coded into 0-9; those coded data, or measured values in some cases, are recorded on data sheets with fixed forms. Eleven items, including cultivar name, evaluation year and others are registered every year in a computer system of the Central Bank together with other items directly related to the characters under evaluation. Under this system, various types of comprehensive information on fruit trees could be easily obtained through the database: in particular, any variety or group of varieties which might have suitable characters for breeding programs could be identified within a short time from among a number of varieties preserved in the Genebank.