Constructing floating offshore wind farms can utilize more stable wind resources in the deep sea

but the wake effect will have negative impacts on wind farm output power. Compared to bottom-fixed wind turbines

the mobility of floating wind turbine can further mitigate adverse effects of wake on wind farm output power by optimizing their positions

and fully consider influences of yaw angle and axial induction factor in the meanwhile. Whereas‌ wake effects create complex interaction relationships that hinder power optimization in floating wind farms

centralized optimization also experiences diminished efficiency as the number of wind turbines rises. Therefore

for reflecting physical mobility characteristics to position constraints

the moveable range of floating wind turbine is firstly characterized. Subsequently

for further discussing influences of mobility on wake relationships on the basis of position constraints

a Gaussian-wake based wake coupling relationship model is proposed for describing wake coupling intensities. Then

for transforming the centralized optimization to the partitioning-based one through constructing the stable digraph based on wake coupling intensities

a floating wind farm partitioning based on label propagation algorithm improvement algorithm is proposed. Finally

simulation and verification are conducted in a regular layout 5×5 floating wind farm with 25 wind turbines. Results show that

compared to only optimizing yaw angles and axial induction factors

the wind farm output power can be raised by 2.443%~11.189% with positions collaboratively optimized. The proposed partitioning-based power optimization method reduces the mean calculation time by 69.776% compared to centralized optimization.