Taking Fujian province as the research object and basing on the CCD data of HJ-1, four vegetation indices were selected which closely related to LAI. The selected indices included NDVI, RVI, mSAVI and EVI. Meanwhile, data synchronized with LAI of Pinus massoniana was obtained, and then the relationship between the indices and the LAI was built. At last, six kinds of the non linear regression models, between the LAI of Pinus massoniana and the selected index RVI that most closely related to the LAI, was established. It was found that the model based on the Richards distribution function is the best and the relative error of the prediction of mean value is 3.1%. Based on this model, the remote sensing inversion of the LAI of Pinus massoniana in Fujian province was implemented.