China is one of the countries with serious concerns about mycotoxin contamination of agricultural food and feed commodities in the world. Mycotoxin contamination leads to a substantial economic loss of grain and oil products coupled with public health hazards; hence, the rapid detection and control of mycotoxin are imminent. The traditional wet chemical detection methods cannot meet the needs of rapid and real-time detection in the process of grain production, supply, distribution, and processing. Even though classical techniques such as HPLC are accurate and sensitive, they have the disadvantage of being time-consuming, entail complex sample preparation, expensive and consume large volumes of chemical reagent. Molecular spectrum is the spectral response produced by the transition between the vibrational or rotational energy levels of molecules, which interprets the structural information of molecules. It can determine the rotary inertia, band length, bond strength, and dissociation energy of molecules, and can be used for the detection of chemical components and properties in samples. The light produced by the transition of the molecules of mycotoxin contaminated grain sample under the excited state is acquired by the photodetector through the optical path system. The spectral intensity and the concentration of the tested substance are underpinned by the Lambert-Beer law within a certain range, which can realize the rapid and quantitative detection of mycotoxin in grain. Compared with the traditional methods of fungal toxins detection, spectral analysis technologies have significant technical advantages of rapid, non-destructive and green. The importance and urgency of mycotoxin detection in grain were analyzed, and then the technical principle and theoretical basis of spectral analysis techniques employed for the detection were introduced. Near-infrared spectroscopy is the vibration caused by the change of electric dipole moment, Raman spectrum responses to the vibration caused by molecular polarization, while the fluorescence spectrum reflects the molecular information with long conjugated structure. Spectral imaging expands from one-dimensional to two-dimensional distribution in detection, and detects mycotoxin quickly and accurately by spectral and feature analysis. This work analyzed the research progress and development trend of different spectral analysis technology, and also exposed the advantages and disadvantages of each technique. The investigation revealed the increasing researchers focus on this research field, and the detection and exploration of grain mycotoxin based on spectral analysis technology, which has become a research hotspot of food safety. Through literature review, it can be found that spectral analysis technology provides a novel approach for rapid screening, qualitative identification, or high-sensitivity detection of mycotoxin in food, but there are still many problems that need to be solved. The applications along with major barriers and limitations of these spectral techniques are discussed, with emphasis on the development of recognition, accuracy and stability. Spectroscopic techniques have the potential to fulfill the need for mycotoxin detection. However, they still require enhancement of theory interpretation, detection scale and accuracy. We believe this review will be an effective guide for rapid detection of mycotoxin in the grain to provide a methodological reference.