Abstract: This study investigates the reduced-order model of wind turbine wakes based on the sparsity promoting dynamic mode decomposition（SPDMD） method using large eddy simulation（LES） results. A comparison is made between the decomposition outcomes obtained from SPDMD and the standard DMD method. The findings demonstrate that the dynamic mode decomposition approach can extract the dynamic characteristics of wake flow, thereby revealing the underlying evolution patterns of wind turbine wakes. The standard DMD method tends to prioritize modes with small scales and high frequencies, whereas the SPDMD method selects large-scale flow features with low frequencies. In comparison to the standard DMD method, the SPDMD approach constructs a reduced-order model of wind turbine unsteady wake fields in a lower-dimensional subspace. By utilizing a smaller number of modes, the SPDMD method improves computational efficiency in the reconstruction and prediction of wind turbine wake fields.