Abstract: Due to the status of cloudy weather, fragmented terrain and diversified crop system in karst mountain region, the conventional remote sensing is difficult to meet the requirement for tobacco growth monitoring. To overcome these problems, we proposed a tobacco growth model monitoring method based on the brightness of the high resolution synthetic aperture radar (SAR) image in karst mountain region. Based on the high resolution SAR image brightness values in karst mountain region, the coupled model of tobacco growth monitoring used regression analysis simulation was established finally. The results from the example analysis and the measured data verification suggested that the established linear regression models between brightness values of SAR and the leaf length and width in the whole growth periods could reflect the linear regression basically, except the one of length in the vigorous growing period. In addition, the tobacco growth condition monitoring was promising for the Yunyan 87 in study area. This method provides a new research thought of large scale and real-time remote sensing monitoring for the modern tobacco agriculture.