As a wide-field microscopy, structured illumination microscopy (SIM) enables super-resolution and three-dimensional (3D) imaging. It has recently received lots of attention due to the advantages of high spatial resolution, short image recording time, and less photobleaching and phototoxicity. The SIM has found numerous important applications in time-lapse imaging of living tissues and cellular structures in the field of biomedical science. Color information is an important physical quantity describing the characteristics of living creatures and reflects the differences in its microstructure and optical property to some extent. Although HSV (hue, saturation, value) color space based structured illumination full-color 3D optical sectioning technique can recover the full color information on the surface of the samples without color distortion. However, for each optical sectioning, three raw images with fixed phase shift are required to calculate the sectioning images by the rootmean square (RMS) algorithm. This will dramatically increase the data acquisition time and data storage space, especially for a large-scaled sample that needs image stitching strategy. The image processing progress operated in HSV color space need to run the RMS algorithm three times in each channel of HSV space for every section, and transform the images between RGB (red-green-blue) space and HSV space twice. This will absolutely extend the data processing time and put forward higher requirements for computer hardware and software for data storage and processing. To this end, in this paper, a fast 3D color optical sectioning SIM algorithm based on Hilbert-transform is proposed. The Hilbert-transform has proved to be a powerful tool in digital signal and image processing and has successfully applied to the SIM. Here, only two raw images with structured illumination are needed to reconstruct a full-color optical sectioned image for each slice. This fast 3D color sectioning method has the advantage of insensitivity to phase-shift error and has better adaptability to noise, high quality color sectioning images can be obtained under the phase-shift error or noise disturbed environment. The image acquisition data are reduced by 1/3 and the color optical sectioning reconstruction time is saved by about 28%, this new method effectively improves the efficiency and speed for 3D color imaging and will bring a wider application range for SIM.