CT-guided needle biopsy is an important minimally invasive procedure to obtain tissue cells from lung nodules for cytological lung cancer diagnosis. The current procedure requires a specially-trained interventional radiologist to manually advance the needle to the target nodule. Accurate sampling depends on the dexterity and consistency of the clinician and the compliance of the patient. Biopsies in patients who have difficulty in holding breath, small nodules or nodules in the lower-third chest are difficult. Long-time remaining still and holding breath degenerate the patient comfort and compliance. Inaccurate needle positioning and insufficient patient compliance may increase the number of needle passes and occurrence of complications. To addresses these problems, we propose a robotic needle biopsy technique which uses a robot manipulator to place the biopsy needle on a target lung nodule according to the patient's respiratory motion pattern, under the guidance of CT imaging. Comparing with manual technique, the robotic needle placement will bring higher operational steadiness, improve accuracy, and reduce procedure duration. With the ultimate goal to create a clinically applicable CT-guided robotic needle biopsy procedure for lung nodules, the goal of this thesis work is to design the robotic needle gripper, a critical component of the robotic needle biopsy system. This work mainly focuses on the mechanical design and analysis of the robotic needle gripper mechanism. Moreover, for patient safety, the design incorporates a set of force sensors to monitor the forces acting on the gripper, which will trigger the release of the needle under serious force imbalance to protect the patient from serious injury, in particularly under sudden cough or large body movement. Our tests on the robotic needle biopsy prototype show that the proposed robotic needle biopsy can achieve high needle placement accuracy on moving nodules.