This article investigates the application of phase-controlled geostatistical inversion in the prediction of fault-controlled hydrocarbon reservoirs, and delves into the prediction techniques for ultra-deep fault-controlled hydrocarbon reservoirs in the Shunbei area. The research team established various types and scales of fault-controlled reservoir geological models and compared the effects of different inversion methods, discovering that controlled geological statistics inversion holds significant advantages in the prediction of fault-controlled reservoirs. In this study, seismic data were used to extract structural tensor attributes to constrain the primary lithofacies in the nested model. Enhanced coherence (AFE) attributes and relative wave impedance attributes were extracted to constrain the secondary lithofacies in the nested model, resulting in a three-dimensional low-frequency model with nested lithofacies at two levels. Leveraging the advantages of nested lithofacies at two levels and step-by-step constraints, the interference from strong seismic horizontal layered reflections was gradually reduced, enhancing the predictability of cave and fracture facies and improving the accuracy of fault-controlled reservoir inversion predictions, making the results more consistent with geological laws. Actual drilling results demonstrate that the predictions are accurate and reliable, with high resolution, effectively guiding well location design and demonstration, and strongly supporting the submission of block reserves. In summary, phase-controlled geostatistical inversion technology provides new ideas and methods for the prediction of fault-controlled hydrocarbon reservoirs, possessing significant practical significance. With continuous technological advancements and further research, the application prospects of this method in oil and gas exploration and development are promising.