This research aims to understand the role images currently play within the design process, in order to develop a classification of image types and reference keywords to construct an electronic image database for professional use in product design. Images play an important role in the design process, both in defining the context for designs and in informing the creation of individual design. They are also used to communicate with clients, to understand consumers, to assist in expressing the themes of the project, to understand the related environments, or to search for inspiration or functional solutions. Designers usually have their own collections of images, however for each project they still spend a significant amount of time searching images, either looking within their own collection or searching for new images. This study is based on the assumption that there is a structure that can show the relationship between the image itself and the information it conveys and can be used to develop the database. A product-image database will enable designers to consult images more easily and this will also facilitate communication of visual ideas among designers or between designers and their clients, thus augmenting its potential value in the professional design process. Also, the value of an image may be enhanced by applying its linguistic associations through descriptions and keywords which identify and interpret its content. Through a series of interviews, workshops, and understanding relevant issues, such as design method, linguistic theory, perception psychology and so on, a prototype database system was developed. It was developed based on three information divisions: SPECIFICATION, CHARACTERISTIC, and EMOTION. The three divisions construct a model of the information which an image conveys. The database prototype was tested and evaluated by groups of students and professional designers. The results showed that users understand the concept and working of the database and appreciated its value. They also indicated that the CHARACTERISTIC division was most valuable as it allows users to record images through their recollection of feelings.