Abstract: Objectives: High-resolution agile optical satellites can continuously observe specific regions or targets, thereby realizing applications such as regional monitoring and target tracking. Existing agile optical satellites are generally equipped with gaze observation capability, while the continuous observation data are geocorrectioned and stabilized to obtain stable and high-quality video data. Nonetheless, owing to the substantial quantity of calculations required during the process, it proves to be challenging to meet the computing and storage requirements on-board, and the processing can only be accomplished on-ground system after the fact, resulting in a long delay for information acquisition. Methods: To address this problem, this paper proposes an object-space-consistency-based real-time stabilization approach for LuoJia3-01 video data, which is based on the data characteristics and the computing capability of embedded on-board hardware. The approach employs real-time geometric positioning to extract and correct the region of interest (ROI) in real time during satellite imaging. Additionally, it employs the consistency of geographical information between image frames to real-time align adjacent image frames to ensure the relative accuracy between sequence frames. Base on this, it generates a sequence of video frames of the ROI with geographical encoding in real-time through the construction of a parallel pipeline of algorithms.Results The experiments are conducted using the actual data and frame rate (6FPS) of LuoJia3-01 satellite, and the results show that this approach can realize the processing speed faster than 6FPS based on the embedded on-board processing hardware. Conclusions: The obtained video frame sequence with geographic encoding and the inter-frame stabilized accuracy is better than 0.328 pixels, which meets the requirements of on-board real-time service.