Variety of color masking technologies have been applied for hardcopies to correct the cross talks caused by unwanted absorption in colorants. Nonlinear color masking technology has been conveniently applied to reduce the reproduction errors on color printers. The color reproduction accuracy is expected to be furthermore improved when the masking matrices are optimized in sub-divided smaller color spaces than in whole color space.This paper proposes a new method to improve the color masking accuracy for hardcopies. Here the input color space is partitioned into numbers of sub-spaces divided by radius, hue angle, chrominance, or others in CIELAB space so that each sub space equally includes the color chip samples inside the printer gamut. Linear or nonlinear color masking functions are applied to each sub-divided space and the masking coefficients are optimized individually by the method of least squares. The new method resulted in the high precision color matching with rms color differences ΔE_ab*(rms) ≈ 1.5 for inkjet printer. As compared with conventional linear or nonlinear color masking methods, the color reproduction errors could be extremely reduced.