College of Geo?exploration Science and Technology, Jilin University, Changchun 130026, China 

 Heavy metal pollution in farmland soil seriously affects crop growth and human health. In order to explore a fast and accurate remote sensing monitoring method for heavy metal content in soil, the Suzi River basin in Liaoning Province was taken as the study area to convert the measured spectral data into multi?spectral data. The authors used the partial least squares regression method (PLSR) to establish a correlation model, and selected the optimal model to apply to the Sentinel??2 images of Suzi River basin in 2016 and 2020. The results show that the comprehensive measured spectrum and multi?spectrum can be used to monitor the content of heavy metals in soil. Each model's R 2 is greater than 0. 50. During the two years, the contents of Cu, Ni, Pb and Zn in the farmland soil increased, and the contents of Cd, Cr and As decreased. The excessive heavy metal elements changed from Cd, Cr and Cu to Cu, Ni, Pb and Zn. Through analysis, the changes of heavy metal contents are closely related to mines, agricultural planting structures, and domestic waste.