It is necessary to stitch small area of images together for large surface analysis while the measurement instrument used with a limited measurement area, e.g. White-light Interferometry (WLI)-based system. A new stitching method is proposed in this paper for diamond grinding wheel surface analysis. The images are captured by a WLI-based system and the 3D images’ stitching requires an overlapping region of 30%~50%. First, two-step intensity correlation matching method is used to obtain several pairs of matched points fast and the RANSAC (Random Sample Consensus) algorithm is adopted to screen them to get exact pairs of matched points. Then the measurement errors are adjusted and a stitched topography is got after data fusion. Experiments show that this method can effectively stitch 3D images of diamond grinding wheel together in less than 4 minutes with a correlation coefficient above 0.9 for two horizontal overlapping regions after adjustment.