Based on outcrop, core, logging, seismic and production data, and the formation of fault-controlled karst reservoirs, the types and characterization of Ordovician fault-controlled karst reservoir architectures in the Tuoputai area of the Tahe oilfield are studied. According to the concept of genetic geologic body, the fault-controlled karst reservoir is divided into architecture elements of four levels, the strike-slip fault impact zone is the level-1 architecture element, the fault-controlled karst reservoir the level-2 architecture element, the fracture-cave zone (which can be further subdivided into dissolution cave, dissolution pore and vug, and fracture zones) inside the fault-controlled karst reservoir the level-3 architecture element, and fillings inside caves is the level-4 architecture element (which can be further divided based on the filling degree and lithologic types of the fillings). Specific characterization techniques have been optimized according to the characteristics of various architecture elements. The zone impacted by strike-slip fault is characterized by seismic coherence and artificial interpretation. Under the constraint of zone impacted by strike-slip fault, fault likelihood (FL) property is used to characterize the outline of fault-controlled karst reservoir. Under the constraint of fault-controlled karst reservoir outline, the internal structures are divided based on seismic texture attribute. Finally, the cavern filling pattern is interpreted based on drilling and logging data. The fault-controlled karst reservoirs can be interpreted in 3-dimensional space by architecture element levels, and the characterization technology combining log and seismic data for fault-controlled karst reservoir has been worked out, which has complemented the development theory and technologies for this kind of reservoirs in the Tahe oilfield.