Soil moisture is one of the important factors affecting agricultural production, and plays a key role in crop growth and development and final yield. However, the waste of agricultural water is widespread, and the utilization rate of water is low. It is estimated that by 2020, China’s irrigation water utilization coefficient will only be It is 0.55, far below the advanced world level of 0.7~0.8. Therefore, the accurate and effective judgment of soil water content abundance is of great significance to agricultural production practice. The detection of soil moisture content has gradually become a research hotspot at home and abroad. Spectral technology can obtain the image information and spectral information of the target simultaneously by utilizing the different characteristic lines of the object, thereby more intuitively expressing the characteristics of the target, and can dynamically and accurately detect the soil moisture content accurately and non-destructively. The technology has greatly promoted the precision, intelligence and modernization of agriculture, and plays an important role in the detection of soil moisture content. This paper reviews the latest literature on soil moisture content detection at home and abroad, systematically discusses the research progress of soil moisture content detection based on spectroscopy technology, analyzes the shortcomings of traditional methods, and expounds the advantages of spectral imaging technology: (1) Real-time; (2) non-destructive; (3) accuracy; and its limitations in the detection of soil moisture content: (1) complex soil structure; (2) insufficient generalization ability; (3) climatic conditions. Three key technologies of spectrum in soil moisture detection are highlighted: (1) Spectral data pretreatment technology, which focuses on the common pretreatment technology principles and effects; (2) Spectral feature extraction technology, which compares common feature spectrum extraction and focuses on the sensitive band of soil moisture. (3) Spectral modeling technology, which focuses on the linear and nonlinear models of soil moisture content detection, analyzes its principle, application range and model accuracy, concludes that the nonlinear model will become the mainstream modeling method of spectral technology in soil moisture content detection. Finally, based on the above analysis, the application prospects and research trends of spectroscopy in the field of soil moisture detection have prospected: Firstly, the generalization ability and robustness of the technology should be improved, and a moisture detection model that can be used for various soil types is established. Secondly, establish a large-area field and dynamically update the soil moisture spectrum database to improve the accuracy of the model.