Development and comparison of the latent fingerprints (LFPs) are two major studies in detection and identification of LFPs, respectively. However, integrated research studies on both fluorescent materials for LFP development and digital-processing programs for LFP comparison are scarcely seen in the literature. In this work, highly efficient red-emissive carbon dots (R-CDs) are synthesized in one pot and mixed with starch to form R-CDs/starch phosphors. Such phosphors are comparable with various substrates and suitable for the typical powder dusting method to develop LFPs. The fluorescence images of the developed LFPs are handled with an artificial intelligence program. For the optimal sample, this program presents an excellent matching score of 93%, indicating that the developed sample has very high similarity with the standard control. Our results are significantly better than the benchmark obtained by the traditional method, and thus, both the R-CDs/starch phosphors and the digital processing program fit well for the practical applications.