Abstract: A wearable wireless health monitoring system for drug addicts in compulsory rehabilitation centers was proposed. The system can continuously monitor multiple physiological parameters of drug addicts in real time, and issue early warning information when abnormal physiological parameters occur, so as to play the role of timely medical practice. In addition, this study proposes a convolutional neural network (CNN)model, which can evaluate the health status of drug addicts based on multiple physiological parameters. Experiments show that the model can be applied to the task of body state recognition in the open physiological parameter data set, and the recognition accuracy can reach up to 100% in a single physiological parameter data set; when the whole physiological data set is used, the recognition accuracy can reach 99.1%. The recognition accuracy exceeds the performance of the traditional pattern recognition method on this task, which verifies the superiority of the model.