Improving the inversion accuracy of soil moisture by optical and passive microwave remote sensing data is an important task to develop quantitative remote sensing. Based on the Soil Moisture Experiment in 2002 (SMEX02) data set, the relationship between the surface soil moisture and the L-band soil emissivity was analyzed. We also discussed the influence of vegetation on soil microwave radiation and deduced a new soil moisture inversion algorithm that takes L-band soil emissivity and Normalized Difference Vegetation Index (NDVI) as independent variables. The SMEX02 experimental data indicates that the correlation between the soil moisture and the L-band soil emissivity decreases rapidly with the increase in NDVI. The verification result shows that the new algorithm developed in this paper has higher inversion precision than the empirical algorithm for the surface soil moisture covered by crop canopy. In relative terms, the inversion RMSE increased from 0.053 to 0.047 for H polarization and from 0.0452 to 0.0348 for V polarization. The R2 variable increased from 0.70 to 0.81 and from 0.79 to 0.86 for H and V polarization, respectively.