Along with the improving of numerical weather prediction models and the application of atmospheric observational data,especially non-conventional sounding data such as satellite and radar,data assimilation methods have made further advancement.Previous data assimilation approaches are reviewed,and the current studies and prospects in the field are described.The data assimilation researches experienced several stages from those early explorations to the subsequent techniques mainly based on experiences(including SCM and nudging),then statistical and the multivariate statistical interpolation(OI,3D-Varand PSAS) methods.Considering the difference between the fact that background error covariance keeps unchanged and the actual situation,the current assimilation techniques take into account the time dimension of sounding data,which are becoming popular internationally and represented by 4D-Var and Kalman filtering techniques.With the development of computer technology,more perfect FDDA methods will play an important role in the future operational prediction.