Carbonate reservoirs in western China have tremendous potential and are the key area for oil and gas exploration in that country. Compared to conventional carbonate reservoirs, discovering and characterizing this kind of reservoir is much more difficult and presents significant challenges. Solutions to the major geophysical problems began with a detailed analysis of exploration issues for such complex carbonate reservoirs and, subsequently, a set of techniques for complex carbonate reservoir prediction was established. These include highly accurate seismic acquisition, prestack amplitude-preserved processing including prestack depth migration, new rock physics models, prestack elastic inversion, seismic multiattribute optimization and integrated prediction, fracture detection using prestack data with limited azimuthal distribution, and fluid identification based on prestack inversion. Additionally, to solve some special problems, diffraction wavefield separation and imaging algorithms were developed for better imaging of caves. Q-migration and nonstationary deconvolution algorithms were also developed for resolution enhancement. Application of these techniques to field data has proven effective, and drilling results show good agreement with our research. Nevertheless, the field is still in its infancy.