A machine behavior recognition method and apparatus, and a device and a computer-readable storage medium, which relate to the technical field of safety protection. The method comprises: acquiring target data to be subjected to recognition and a machine behavior recognition model, wherein the machine behavior recognition model comprises a principal component analysis layer, a gradient descent tree model layer, a random forest tree model layer and a logistic regression model layer (S101); inputting the target data into the principal component analysis layer for processing so as to obtain a principal component characteristic of the target data (S102); inputting the principal component characteristic into the gradient descent tree model layer for processing so as to obtain a first machine behavior recognition result of the target data (S103); inputting the target data into the random forest tree model layer for processing so as to obtain a second machine behavior recognition result of the target data (S104); and inputting the first machine behavior recognition result and the second machine behavior recognition result into the logistic regression model layer for fusion processing so as to obtain a machine behavior recognition result of the target data (S105), wherein the machine behavior recognition result can be uploaded to a blockchain for storage. By means of the method, the recognition accuracy of a machine behavior is improved.