Abstract: As an important problem restricting agricultural development,soil salinization not only causes great damage to the local ecological environment,but also reduces agricultural food production,which affects the sustainable development of agricultural economy.This study selected typical soda saline-alkaline soil areas in the Songnen Plain as the research area to compare and analyze the spectral response to soil samples with different salinity levels at different heights.Specifically,the spectral reflectance of five experimental heights including 3 cm,4 cm,5 cm,6 cm,and 7 cm were accurately obtained.Correlation analysis was then performed between the soil electric conductivity(EC) and the spectral parameters including the reflectance,the first-order and the second-order transformation results of reflectance.After that,four sensitive bands with high correlation coefficient were selected to establish multiple linear regression model and partial least squares regression model.The results indicated that the predictive ability of the partial least squares regression model was better than the multiple linear regression model,and the correlation coefficients between EC and spectral parameters at different measurement heights were also quite different.In general,the partial least square model established by the result of the second-order differential transformation of the reflectance at the measurement height of 5 cm demonstrated the best fitting ability on soda saline-alkali soil samples with R²=0.91,RPD=3.08 and RMSE=0.40,respectively.The purpose of this study was to select the best measuring experiment height through outdoor spectral experiment,to improve the prediction accuracy of soil salt content and therefore to provide feasible suggestions for improving soil salinization in Songnen Plain.