Abstract: In order to test the accuracy of the automatic instrument for measuring the fuels moisture content, the obtained moisture content data are compared with the manual measured moisture content data, analyze the causes of errors combined with environmental factors and the working principle of the instrument, and find a correction method to improve the accuracy of the instrument. In this study, three kinds of dead fuels on the ground surface of Harbin were taken as the research object, and a new generation of automatic measuring instrument of moisture content was set up to obtain moisture content data and meteorological data, and compared with the field measured moisture content data. The instrument data were corrected by establishing linear and nonlinear regression models. Finally, the extrapolation of the model was tested and the calibration method was determined. The results showed that（1）The difference between the instrumental data and the measured data was mainly due to the systematic error, which was caused by the water or other impurities attached to the net pocket. This was mainly related to the complex environment in the forest and the working principle of the instrument, which needed to be corrected.（2）Meteorological factors also had a certain influence on the significant difference between instrument measured moisture content data and manual measured data.（3）Among the two calibration models, only the linear regression model had good correction effect and extrapolation, and the moisture content data measured by the instrument can be corrected, and the corrected data can meet the accuracy requirements. There was a certain error in the automatic measuring instrument of fuel moisture content, which was significantly different from the manual measured data. It can be corrected by the linear regression model, and the corrected instrument data met the accuracy requirements of the measurement, which improved the accuracy of the instrument in measuring moisture content data, and can provide important technical support for fast and accurate fire risk prediction in the future.