Affective values are critical factors for manufacturing in Japan. Kawaii, an affirmative adjective that denotes such positive meanings as cute or lovable, has become even more critical as an affective value and plays a leading role in the worldwide success of many products, such as Hello Kitty and Pokemon. Based on this success, we believe that kawaii will be a key factor in future product design. In our previous research, we proposed models of kawaii feelings for spoon designs and extracted the attributes of such designs and constructed models using the Support Vector Machine (SVM) algorithm. In this research, we used the Deep Convolutional Neural Network (CNN) algorithm because it can perform classification using images as input and studied the kawaiiness of cosmetic bottles. Then, we evaluated the candidates of effective attributes with our model to increase the kawaiiness of cosmetic bottles. Finally, we clarified the relationship among kawaii feelings, attributes, and eye movement indexes obtained from our previous research, and the prediction results of our constructed model. Our results clarified the effective attributes for increasing kawaiiness and the effectiveness of our constructed model to evaluate the kawaiiness of cosmetic bottles.