At present, it is difficult to accurately detect interharmonics at an adjacent frequency and in signals with transient disturbances in a power system. For the above two types of interharmonic components, first, a matrix pencil algorithm is used to detect the steady-state signal with adjacent interharmonics in noise. Secondly, for the interharmonics in the transient disturbances signal, the Singular Value Decomposition (SVD) and matrix beam algorithm combined detection method is proposed. This uses SVD to accurately locate the moment when the signal is abrupt, decomposes the transient disturbance signal into multi-segment stationary signals and then uses the matrix pencil algorithm to calculate the modal parameters. The accuracy of the above method is verified by the detection of Matlab simulation data and the sub-synchronous oscillation simulation signal of the fan. The method is used to detect the harmonics and interharmonics in the stator current when the output power of a wind farm double-fed fan is changed. The results are consistent with the theoretical analysis and the methods have strong practicability. This work is supported by the National Natural Science Foundation of China (No. 51777066).