Abstract: There are rugged surfaces and complex subsurface structures which lead to raw seismic data with poor quality in Baka area of Northern Turpan-Hami Basin. Especially for the main parts of subsurface structures, there are seriously steep formations and abundant faults. So it is difficult for seismic migration and to determine the subsurface structures. And the reservoirs feature in low permeability, low porosity and inconspicuous geophysical response. So there are few reservoir prediction and identification methods. The above factors restrict the progress of oil and gas exploration of Baka area. Technical series have been developed based on research, such as seismic imaging, structural interpretation and reservoir prediction techniques. And the key innovation techniques include model constrained integral static correction method for rugged surface, "six division" prestack noise attenuation method and floating datum pre-stack reverse time migration method. On the basis of combining research and production, the new research results could more accurately describe the subsurface faults and structures, and the overall imaging effect has been improved effectively. Application of new research results helps us determine the structural types again, further realize the faults characteristics and formation distribution rules, find and determine some new traps. On this basis, research reservoir prediction is conducted and several beneficial drilling targets are chosen

Key words: constraint static correction, "six division" noise attenuation, reverse time migration, effective reservoir prediction, fracture prediction