In lumbar puncture surgery, compared with the conventional methodologies like computed tomography and magnetic resonance imaging, ultrasound imaging offers the advantages of being low cost, no radiation and real-time image generation. However, the use of ultrasound equipment in lumbar puncture involves a cumbersome and time-consuming process for the subjective imaging of the overall structure of the lumbar spine in order to determine the exact puncture point and path. Meanwhile, the robotic arm puncture system has the advantages of high precision, good stability and simple and efficient operation. As a result, robotic-assisted ultrasound scanning is valuable for the assessment of a puncture path in spinal tap surgery. In this pursuit, based on the official URSDK development package for a robot arm and the Transmission Control Protocol/Internet Protocol, the system proposed in the present study involves a program to control the robot arm to clamp down onto an ultrasonic probe to enable automatic scanning and acquisition of images. A three-dimensional reconstruction program based on the visualization toolkit was designed, and a lumbar spine experiment was conducted with this system. A total of 136 two-dimensional ultrasound images were collected in the lumbar spine model experiment by enhancing contrast of and denoising the original ultrasound images, and a linear interpolation algorithm was used to perform the three-dimensional reconstruction of the lumbar spine model. The reconstructed structure was defective, but the location of the spinous process gap was determined with the sagittal and coronal images. The feasibility of the system was verified by the reconstruction results, which can provide a reference for determining the puncture point and path planning in the lumbar puncture surgery.