A system to support the design of cosmetic containers is discussed. The proposed system aims to create a design plan by cooperating with the designers. Cosmetics can be classified as emotional products because users tend to select them based on the images and the persona of the products. Cosmetic product images depend on the container design, the graphic design, the brand design, etc. The design process for cosmetic containers is important, and designers should understand how users respond to containers. However, it is usually difficult to quantify users' feelings about products. With the proposed system, the designers refer to evaluations from multiple points of view gathered by the system in making their design plans. Two methods of estimating users' evaluations are introduced in the system. The first method uses regression analysis, and the second method uses artificial neural networks (ANNs). Questionnaires about users' impressions of cosmetic containers were given, and the results were used to identify parameters of a regression formula and teach values of ANNs. In addition, the system was used for the actual design process of a premium cosmetic container, and evaluations of the product were investigated. As a result, 49% subjects evaluated it as a luxury item.