During clinical liver biopsy surgery, the high-precision and low-trauma advantages of the biopsy robot (BR) are hindered due to the interference of respiratory motion (RM). To mitigate the interference, simultaneously reduce dependence on the skill of surgeons, and enhance the autonomy of the robotic system, an autonomous liver biopsy robotic system (ALBRs) is designed. A spatial pose-constrained calibration algorithm is proposed to achieve high-precision identification of postural relationships among system components. A target pose-localization and follow-up algorithm is developed to accurately locate the target and follow up its dynamic posture under RM. Furthermore, an online liver biopsy strategy is designed to enable autonomous biopsy operations. In addition, an RM simulation device (RMSD) is developed, which adjusts motion parameters in real time based on authentic human respiratory data to simulate liver biopsy conditions under RM. Validation experiments demonstrate the high feasibility of RMSD. An innovative experimental method verifies the calibration algorithm, achieving maximum, minimum, and average errors of 0.42, 0.12, and 0.23 mm, respectively. Extensive experiments, including animal- and model-based tests, are conducted to comprehensively evaluate the system. Animal experiments yield distance and angle errors of 1.60 mm and 0.32 degrees, while simulated membrane experiments result in errors of 2.13 mm and 1.38 degrees. These results confirm that the proposed system and algorithms exhibit commendable robustness and precision, meeting the clinical requirements for liver biopsy surgery.