Anisotropic migration has become a routine practice for fracture prediction because of its remarkable capability of suppressing effects of geologic structures while preserving anisotropy induced by fractures. The quality of input seismic gathers is of the first priority among other factors for fracture prediction. Azimuth-sectored migration loses considerable accuracy and reliability because of its coarse and inappropriate manner in dealing with seismic data. A full-azimuth anisotropic migration in surface offset domain provides better efficiency in using data information. Furthermore, full-azimuth anisotropic migration in local angle domain (subsurface angle domain) allows for accurate and reliable exploitation of seismic wavefield information. Finally, influences of the above three anisotropic migration methods on fracture prediction are prescribed with a field example in which validations with formation micro-imager (FMI) information demonstrate that full-azimuth anisotropic migration in local angle domain is the state-of-the-art technique for recovering true properties of subsurface fractures.