Low amplitude structure developed in Jurassic Yanan Formation in the southern section of Tianhuan Depression.Due to the influence of complex tectonic evolution of the conterminous southwest margin，it is difficult to reveal the regularity of hydrocarbon accumulation in the low amplitude structure. So exploration deployment is difficult. In order to improve the prediction effect of low amplitude structure of the Yanan Formation in the southern Tianhuan Depression that is covered by 2D seismic data，tomographic static correction technology was optimized to solve the problems of long and short wavelength imaging，and therefore the precision of low amplitude structure mapping can be improved. After that，reverse linear noise suppression technique based on data reconstruction was used to reprocess the data which have problems of strong interference and low signal-to-noise ratio in loess tableland area. On the basis of fine processing，by using subtle velocity field，variable-velocity and small grid of large-scale mapping，the“fasle-high”caused by high-speed conglomerate-interlayer was removed and the structure can be mapped correctly. Finally，the trend surface analysis was utilized to highlight the local structure heights. Nineteen low amplitude structures were determined based on comprehensive use of the integration of data process and interpretation mentioned above. It has been proved by drilling that the low amplitude structure prediction technique has a good effect on low amplitude structure prediction in area that is covered by 2D seismic data.