This study attempts to prove the possible negative effect phenomenon of excessive fit, i.e. match of CSR (Corporate Social Responsibility) activity or campaign with corporate image and related objects. Many studies about fit effect of CSR have proved and asserted that the more CSR activity matchs with corporate image, the more effect it brings. It has generally been known there is a positive relationship between degree of fit and attitude (also intention to purchase, etc.) toward corporate. However, some recent studies showed the fitness does not necessarily assure a good effect. This study, criticizing that past studies only used single or uni-dimensional variable approach, elaborated more in detail upon the variables and research methods, including statistic analysis, thus drawing a more valid conclusion. The 4 multi variables used were functional fit, image fit, target fit, size fit. The researchers expected an inverted U typed curve or elliptic limited curve between degree of CSR-fit and attitude toward corporate. In addition, the attitude the respondents had toward corporate before the CSR act was included as mediating variable. The curve-linear regression analysis was used instead of linear regression as statistical tool. The results were as follow: There found inverted typed curves or elliptic limited curves when functional fit, image fit and target fit were used as independent variables. The size fit yielded exceptional result, as it showed a just linear relationship. These negative effects were shown more distinct among the respondents who had a more negative pre-attitudes toward corporate.