Spacecraft remainders online detection is affected by insufficient marker data. In this paper, the physical characteristics of remainders generation were studied, a remainders model in spacecraft digital twin systems was established, and a cross-domain adaptive spacecraft remainders detection method with digital twin enhancement was proposed. This method was augmented by digital twins to acquire real-time spacecraft data. It then combined similar structural historical marker data and applied a cross-domain adaptive approach to assist current online reasoning. In addition, a new type of cross-domain adaptive approach model was proposed to better mine prior knowledge from complex tasks through a shared network structure and gating mechanism. The model realizes a combination of cross-domain adaptive techniques and digital twins for more efficient, accurate and real-time prediction. This method can comprehensively detect the remainders states of different components of spacecraft.