Abstract: To address the issues of existing moxibustion robots relying on manual guidance for acupoint recognition, the separation of acupoint identification and moxibustion therapy, and the complexity of multi-point acupoint operations, this paper designs an automated moxibustion robot based on multi-point acupoint recognition. The robot consists of a base, a six-axis robotic arm, an end-effector moxibustion box, an RGB-D camera, and a control system. The control system employs image measurement and recognition algorithms to accurately acquire the coordinates of acupoints marked by therapists. These coordinates are then transformed into the robotic arm's base frame using a hand-eye calibration matrix, enabling automatic acupoint recognition and motion control of the robotic arm. Additionally, the system supports automated path planning for multi-point acupoints, personalized settings for moxibustion duration and temperature, and the simulation of various moxibustion techniques, including suspended moxibustion, circular moxibustion, and sparrow-pecking moxibustion. The system achieves a positioning error of ≤5 mm for multiple acupoint markers. This study effectively resolves the challenges of automatic multi-point acupoint recognition and the simulation of diverse moxibustion techniques, demonstrating broad market potential and application prospects.