The underline idea of gray component replacement (GCR) is to replace a mixture of primary colors (cyan, magenta, and yellow) by black. The conventional algorithms of GCR are mainly based on the concept of equal-tone-value-reduction or mixing equal amount (tone value) of primary colors generating gray, which in turn can be represented by the same amount of black. As the inks used are usually non-ideal, such a replacement can result in significant color deviation. In practice, only partial GCR is applied. We proposed an algorithm of maximal GCR based on color matching, i.e. the black is introduced in a way that preserves the color (before and after GCR). In the algorithm, the primary with the smallest tonal value is completely removed while the other two are reduced and the black is added according to the color matching calculations. The algorithm is verified by colormatching computations, for a large number of target colors. It shows that all the target colors can be matched with negligible color deviations (ΔE smaller than unity). Measurements of test prints of 90 colors which cover broad range of C,M,Y compositions, demonstrate an average color deviation ΔE=2.56 that is slightly bigger than the so called just noticeable difference (JND) and significantly smaller than that of the conventional GCR, ΔE=9.75.