Detection of surface defects is critical in quality control of reflective optics. In this note, we propose a new surface defect detection method for reflective optics using the normalized reflectivity, which is calculated from the signal intensity of a chromatic confocal surface profiler. This detection method first scans the surface to acquire signal intensity data and then models the intensity data to calculate the normalized local reflectivity map. The reflectivity map is further processed by threshold segmentation to extract defects from normal areas. Measurement experiments on an Al-coated concave reflector with artificial defects were carried out to demonstrate the feasibility of the method. This detection method can provide existing optical surface profilers with defect detecting capabilities without extra equipment.