To modernize traditional Chinese medicine (TCM) and improve the efficiency of Tuina therapy in the rehabilitation of cervical spondylosis. This study aimed to solve the problem of the efficacy of cervical spine Tuina therapy, which lacks objective evaluation standards because it is difficult to provide effective feedback for massage robots. A cervical spine Tuina efficacy assessment model built with functional nearinfrared spectroscopy (fNIRS) data was proposed to evaluate the performance of a multi-information fusion TCM massage robot. The robot combined visual guidance technology, a multi-degree-of-freedom robotic arm, fNIRS, and TCM Tuina techniques to provide a scientific auxiliary diagnostic and therapeutic method for TCM Tuina therapy. The feasibility and effectiveness of the robot were verified through two tests: comparing the subjective scale with the objective fNIRS data and comparing the feelings of the physician's massage with the robot's massage. The accuracy of the efficacy assessment model was greater than 95%.