The invention discloses a wide area measurement based power distribution system network anomaly detection method, which comprises the steps of S1, installing a plurality of phasor measurement unit detection points in power distribution network lines; S2, respectively performing parameter initialization on the starting time of sampling, the ending time of sampling, the current sampling time, the time sequence window length and the sliding time interval; S3, solving an effective value for current data sampled by a phasor measurement unit, performing differentiation according to a time interval M, and forming a sampling matrix X according to a time sequence; S4, performing standardization processing on the matrix X to acquire X, and then solving a sample covariance matrix Z of the X; and S5, solving a characteristic root [lambda]Z of the matrix Z, performing spectral distribution analysis on the characteristic root [lambda]Z, acquiring spectral distribution in a complex plane, calculating an absolute value of (rdet-r0)/r0, and comparing the calculated absolute value with a threshold limit K0. According to the method, abnormal signals can be effectively detected, the sensitivity is greatly improved compared with that of a traditional method, the detection time is obviously ahead of anomaly occurrence time, and the method has important significance for analysis and processing for network anomalies.