Abstract: With the process of urbanization and industrialization, heavy metals enter the ecological environment through various pathways and accumulate in large quantities in the soil, causing potential risks to soil environmental health. In recent years, with the development of energy dispersive X-ray fluorescence spectroscopy (ED-XRF), the detection limit of the instrument has been significantly reduced, and it can be effectively applied to the detection of various heavy metals in soil, and is gradually becoming an effective tool for determining the concentration of heavy metals in the soil environment. However, the complexity of the soil matrix and the limitations of the instrument itself will lead to problems such as low accuracy and precision in the determination of target heavy metals by ED-XRF, such as the type and particle size of soil samples and the noise in the instrument and the environment; there are also certain difficulties in the establishment of quantitative analysis model methods, which are still not well applied to laboratories and other environments with high data quality requirements. In this paper, we summarize the application and research progress of XRF in the field of soil heavy metal detection, explore and analyze the influence of different soil sample states and detection conditions on the detection accuracy of ED-XRF instruments, sort out the main pretreatment methods of ED-XRF spectroscopy and the establishment process of quantitative analysis models, and introduce and analyze the application potential of ED-XRF in evaluating the effectiveness of soil heavy metals. At present, the soil sample preparation procedures applied to ED-XRF detection have been relatively mature, and the opinions of scholars on how to reduce the impact of soil matrix effects on the operation methods are relatively unanimous, that is, to use dry samples, low particle size or pressed soil samples as much as possible, and there is still some room for optimization in the processing and analysis of detection data. Therefore, the current mainstream focus is on combining the advantages of different algorithms for the preprocessing analysis of ED-XRF spectroscopy and the establishment of quantitative analysis models to improve the accuracy of ED-XRF detection. At the same time, ED-XRF detection studies on different types of soils only highlighted the differences in the phenomenon, and did not clearly explore the underlying mechanisms. In the future, it is important to continue to explore the causes and magnitude of different matrix effects in various types of soils, and to optimize and improve the quantitative analysis model of ED-XRF spectroscopy; the prediction of the effectiveness of heavy metals in soil through multiple regression analysis is also a field that scholars should focus on.