The work aims to improve the mapping effect of color gamut mapping technology in non-uniform space during the cross-media process. An out-of-gamut mapping algorithm based on the hue tone curve model was proposed through the non-uniformity analysis of CIELAB color space. In this algorithm, these colors were divided into seven categories (red, green, blue, black, pink, yellow and gray) with the same hue and different saturations. Through the space transformation of these seven categories of colors from CIERGB to CIELAB, the spatial curve of seven hues could be obtained by connecting all the scattered points according to certain color sequence to predict those color points with the same hue and different saturations. Based on the method that predicted colors according to the spatial curve, the color points whose hue constant mapping was out of the device gamut boundary were kept. Through the subjective evaluation, the four digits of HPMINDE and SGCK algorithms were 0, the mean value of the new algorithm was 0.8, and the median was 1. Through the objective evaluation, the average color difference was calculated by HPMINDE, SGCK and the new algorithm, which respectively were 12, 11 and 10. It was clearly showed that the color difference of the new algorithm was much smaller. Compared with the HPMINDE and SGCK algorithms, the proposed algorithm has the superior characteristics to keep the hue unchanged and achieve better visual effect of mapping results, and its vision performance is preferable for observers.