Long Due to the influence of wake flow between wind farm units, some wind farms have actual operating conditions lower than the design indicators, resulting in significantly insufficient power generation capacity. This article proposes a wind farm wake collaborative control method based on a double-layer architecture model,which improves the energy capture efficiency of the wind farm area and reduces the impact of wake between units. The method combines DBSCAN clustering and wind direction based dynamic clustering algorithms to group wind farm units, and divides the areas with significant wake impact between units into the same cluster. Performing wake collaborative control for wind turbines in the same cluster reduces wind turbine dimensions and computational difficulty, and improves the efficiency of wake control algorithms. This method is not based on complex wake models and can be quickly applied online.