In the absence of physical movement, the dynamic process of rehearsing specific actions in the mind based on memory is called motor imagination (MI). Motor imaging brain-computer interface (BCI) refers to communicating with a computer or controlling external devices by imagining specific movements in the brain, while the Internet of health things technology helps to convey the rehabilitation of the limbs of patients with paralysis and stroke in time. Studies have shown that with the increase of motor imaging tasks, the classification accuracy rate will also decrease. It is necessary to study the construction and security of sports injury rehabilitation systems under the Internet of health things. Based on the study of the feature extraction algorithm, this paper starts with the infusion monitoring in the clinical medical care work, and through the construction of the Internet of health things smart care management system. It could provide application modules such as "infusion monitoring", "disease call interconnection", and "PDA interconnection" to realize the closed-loop path management of the whole process of basic nursing work in the ward. The sports rehabilitation training system receives the EEG data collected by the UE-16B EEG amplifier through the socket, and calls MATLAB to perform feature extraction and classification of the EEG data, and feedback the processing results to the sports rehabilitation training system and subjects. On this basis, the system has designed and completed a non-feedback training module, a feedback training module and related BCI game modules to promote users to perform sports imagination training and successfully complete sports rehabilitation training.