It is very important to fully understand the relationship between ink amounts placed on the paper and resultant colors for a dye-based inkjet printer (IJ). In the field of printing, many predicting models such as the Neugebauer model (NM), the Yule-Nielsen Neugebauer model (YNNM), and the Cellular Neugebauer model (CNM) have been proposed. However, it is very difficult to precisely predict the resulting printed colors as reflectances or tristimulus values from ink amounts because of the nonlinear relationship between them.In the previous study, the CNM only achieved sufficient prediction accuracy among a variety of predicting models, when the set of primaries was 5^N (N is the number of colorants used). In this paper, it is necessary to print and measure many primaries for prediction of reproduced tristimulus values. For example, 6-colorant prediction needs 5⁶ = 15,625 primaries.The purpose of the present study is to further decrease the number of Neugebauer primaries that the CNM needs to predict with sufficient prediction accuracy. Further, the novel approach is to propose by taking into account dot gain for the present-day subtractive color IJ.