The invention relates to the technical field of intelligent power grids, and particularly relates to an intelligent power grid communication anomaly detection method based on a time sequence. The method comprises the following steps: for any normal communication network flow, determining a sequence position of each element in a period through a period overlapping mapping based on a period extraction method of discrete Fourier transform so as to determine a periodic time sequence relationship of each element in the normal network flow; obtaining the periodic time sequence relationship of each element in the to-be-detected network flow, and comparing the periodic time sequence relationship with the periodic time sequence relationship of each element in the normal network flow, so that anomaly detection of network communication is realized, and the method has very important significance for monitoring and detection of a strong periodic intelligent power grid control flow.