In this paper, we estimated the sensitivity information received from the image, and examined the possibility of the application of the image retrieval. We analyze the sensitivity information received from the landscape image by the based on Semantic Differential evaluation experiment. As the result, four meaningful factors are obtained. Then, by using the multiple regression and potential interpolation, we estimated the value of sensitivity information factors from image characters which are average value of the CIE L^*a^*b^* color space, color histgram, sensitivity information of color and texture characters. Furthermore, we construct the image retrieval system based on the sensitivity words. This system utilize the sensitivity information as the indexing.