Abstract: The inability to obtain surface roughness factor in large scale area is the difficulty in microwave inversion of soil moisture. In this regard, from the perspective of the advantages of synergetic microwave and optical remote sensing data in soil moisture monitoring to improve the inversion accuracy of soil moisture, we studied a remote sensing inversion method for soil moisture in large areas of surface roughness parameters. The impact of vegetation to the backscattering was removed by applying the water-cloud model based on GF-3 and Landsat8 remote sensing data. Then, the backscattering, the incident angle, PDI, TVDI and NDWI index were used together as the model input parameters. The soil moisture inversion model was under 2 polarization modes (HH and HV). The results showed that: R 2,IA ,RMSE were used to compare the experimental results of the two models in farmland of Jianghan Plain, the inversion accuracy of soil moisture under HH polarization was better than that of HV polarization mode (HH polarization method:R 2=0.6864,IA =0.8895,RMSE =6.979%). In space distribution, the areas with higher temperatures and lower vegetation coverage had lower soil moisture, such as the north of Zhongxiang, the northeast of Jiangling, the east of Jingmen, the west of Xiantao, and the northwest of Tianmen; the areas with lower temperatures and higher vegetation coverage had higher soil moisture, such as Jianli and the southern part of Tianmen.