Fungus is one of the primary factors endangering the safety of grain storage. The rapid detection of fungi on stored grain in early stage is an effective measure to prevent and control the fungal multiplication and ensure food security. In 2018, an industry standard “LS/T 6132 Inspection of Grain and Oil—Storage fungal examination—Enumeration spores of fungi” was promulgated and implemented in the grain industry. In this study, we developed an automatic detector for the detection of fungi in grain storage to promote the application of this industry standard. During the development of the detector, we built a fungal spore image library based on a large number of stored grain fungal spore pictures, and developed a fungal spore image recognition software using neural network algorithm. By optimizing the auto focusing algorithm of the microscopic imaging system, the fungal spore image under the microscope can be automatically focused and photographed. And the image recognition software was used to recognize and count the spores of the stored grain fungal automatically. This detector can realize the automatic detection of fungi on stored grain and reduce the probability of mistaken in personnel operation and identification.