The invention discloses a wireless unbound human body behavior detection algorithm. The objective of the invention is to identify human body behaviors through analyzing the different change patterns of channel state information. With the algorithm adopted, high recognition accuracy can be realized, and at the same time, convenience and safety can be realized. According to the technical schemes of the invention, a wireless transmitting end establishes a wifi field; when a user walks or does a certain action in a wifi coverage range, specific influences will be exerted on a wifi channel; and a wireless receiving end is adopted to receive wifi signals, calculate the CSI value of the action, and extract channel change features. According to the algorithm, based on different change characteristics generated for the wifi channel by the movement or action conduction of the person in the wifi field, the different change patterns of the channel state information are analyzed; feature extraction and classification matching algorithms are adopted to carry out behavior identification; the different change patterns of human body behaviors and the channel are combined; and therefore, the human body behaviors can be identified through the features of the wifi channel.