Abstract: It has been proved by theory and practice that the parameters of regression model can not be estimated by the least square method in the design of regression saturation and that the experiment plan designed by regression saturation can not be evaluated by the D-optimality. As a result, the D-optimal design of regression saturation is not existed.