A variety of color calibration technologies has been developed for input/output devices. Linear or nonlinear matrices have been conveniently applied to correct the color filter's mismatch in scanner or suppress the cross talks of colorants in printer. The color matching errors are furthermore reduced when the nonlinear matrices are optimized in subdivided smaller spaces than in an entire color space. This paper proposes a new method for partitioning the color space into subspaces to improve the accuracy. Linear or nonlinear color correction functions are applied to each subspace and the matrices are optimized individually. The new method resulted in the high precision color matchings with rms color differences ΔE_ab *(rms) ≤ 0.5 for flat bed scanner and ΔE_ab *(rms)≅ 2.0 for inkjet printer. The accuracies are approaching to the measurement errors in scanner and mechanical stabilities in printer.