Technical Abstract: Even though cherry tomato is one of the major vegetables consumed in the fresh-cut market, its quality evaluation process has been dependent on simple size- or color-sorting techniques, which currently is inadequate for meeting the increased consumer demand for high quality and safety products. Of the various quality evaluation processes, detecting the defect cherry tomatoes is one of the most crucial processes since cuticle defects could be high potential harbor sites of pathogenic microbes which may cause deleterious consequences to consumer health. In this study, multispectral fluorescence imaging technique is presented as a desirable tool for non-destructive detection of defect cherry tomatoes compare to multispectral reflectance and conventional color imaging. A noteworthy observation in the use of fluorescence images was that the fluorescence intensity of the areas of cuticle cracks was immensely higher than that of the sound surfaces in the blue-green spectral region, which could be useful as a sensitive classification tool for the detection of cracking defects on cherry tomatoes. Simple ANOVA classification analysis as well as principal component analysis was explored to investigate optimal fluorescence wavebands. The results illustrate that the fluorescence spectral imaging technique shows high feasibility for detecting defects on cherry tomatoes. A multispectral fluorescence image linearly combined with a pair of selected wavebands based on the result of principal component analysis was capable of detecting defect cherry tomatoes with above 98% accuracy. The multispectral fluorescence imaging technique showed good potential for discriminating between defect and sound cherry tomatoes. The detection algorithm investigated in this study could be used for developing on-site and real-time multispectral systems for quality evaluation of cherry tomatoes in postharvest quality processing plants.