Color stimuli with different primary colors will arouse different color perceptions, and color matching accuracies calculated by different color matching functions are also different. To investigate the influences of the peak wavelength location of primary colors from self-illuminating equipment on such differences, we selected one 3-channel LED panel (R1G2B1) and one 6-channel LED panel (R1R2G1G2B1B2). The color displayed by the 636 nm-524 nm-452 nm combination in the 3-channel LED panel was defined as the target color, and seven corresponding combinations (including one combination of the same spectrum) were generated after the RGB single channel and double channel of the 6-channel LED panel were changed. Thirty-nine observers (aged from 19 to 24) with normal color vision were organized for the color matching experiments of white color, and 357 sets of data were collected in the visual experiments. Four sets of color matching functions were employed, including CIE1964, CIE2006, Sarkar2, and BIGC17 (the latter two are based on reflective color sample pairs and suitable for young observers). Their performances were evaluated with Δ(u', v') calculated by CIE1976-u'v' chromaticity. The results indicated that regarding the color matching accuracy and the observer variation calculated by CIE1964 and CIE2006, the peak wavelength positions of the red channel, blue channel, and green channel had increasing influences, and the differences from the same-spectrum combination results were also in an ascending order. Sarkar2 and BIGC17 outperformed CIE1964 and CIE2006 in some combinations.