A neural network based approach for female eye glasses fitting suggestion is addressed in this article. The experimental results are analyzed by two stages of back-propagation neural network, which first categorize the Taiwan female faces, and then establish the relationships between adjective image words and female faces with different glasses. A female glasses fitting model was then developed based on these database. With the aid of the model, a glasses store or an electronic store can generate professional fitting opinions to its female customers after a practical ophthalmic dispensing. Although the system is mainly designed for a Taiwan female face, the method can be used to develop for other countries or products.