SUMMARYCommon-conversion-point binning associated with converted-wave (C-wave) processing complicates the task of parameter estimation, especially in anisotropic media. To overcome this problem, we derive new expressions for converted-wave prestack time migration (PSTM) in anisotropic media and illustrate their applications using both 2D and 3D data examples.The converted-wave kinematic response in inhomogeneous media with vertical transverse isotropy (VTI) is separated into two parts: the response in horizontally layered VTI media and the response from a point scatter. The former controls the stacking process and the latter controls the process of PSTM. The C-wave traveltime in horizontally layered VTI media is determined by four parameters: the C-wave stacking velocity V C2 , the vertical and effective velocity ratios γ 0 and γ eff , and the C-wave anisotropic parameter χ eff . These four parameters are referred to as the C-wave stacking velocity model. In contrast, the C-wave diffraction time from a point scatter is determined by five parameters: γ 0 , V P2 , V S2 , η eff and ζ eff , where η eff and ζ eff are, respectively, the P-and S-wave anisotropic parameters, and V P2 and V S2 are the corresponding stacking velocities. V P2 , V S2 , η eff and ζ eff are referred to as the C-wave PSTM velocity model. There is a one-to-one analytical link between the stacking velocity model and the PSTM velocity model. There is also a simple analytic link between the C-wave stacking velocities V C2 and the migration velocity V Cmig , which is in turn linked to V P2 and V S2 .Based on the above, we have developed an interactive processing scheme to build the stacking and PSTM velocity model and to perform 2D and 3D C-wave anisotropic PSTM.Real data applications show that the PSTM scheme substantially improves the quality of Cwave imaging compared with the DMO (dip moveout) scheme, and these improvements have been confirmed by drilling.