The embodiments of the invention provide a human body fall detection method and device. The method comprises the steps of: acquiring a target image; performing human body detection on the target imagevia a target detection network to determine whether the target image contains a human body; when determining that the target image contains a human body, performing fall identification via a convolutional neural network to determine whether the human body in the target image is in a fall state. The method performs analysis processing based on acquired single-frame target images instead of streaming video, identifies images containing human bodies by using the target detection network based on a target detection algorithm, and performs classification identification on human body states in thetarget images by using the convolutional neural network based on a classification algorithm to identify the states of the human bodies in the target images, thereby solving the technical problems of low accuracy and poor efficiency of human body fall identification in the prior art and achieving the technical effect of identifying a fall state accurately and efficiently.