The purpose of this study is to identify a ‘kansei’ space for ‘kandoh’ evaluation model that considers individual characteristics. We performed two experiments that attempted to construct such kandoh models for a younger age group and a mature age group. In the first experiment, 69 suitable words for evaluating the image quality were selected from 150 Japanese words to express kandoh. In the second experiment, using the selected kandoh words, the kansei space for each age group was identified and analyzed by individual differences in scaling, multi-dimensional scaling, and cluster analysis. Based on the results, we observed differences in the interpretation of the words between age groups. Thus, the constructed kansei spaces were shown to also have differences depending on the age group.