Abstract: The offshore wind turbine presents the trend of large-scale development, and the voltage level of the power collection system changes with the change of single unit rated capacity. The fuzzy c-means clustering algorithm（FCM algorithm） is used to optimize the topology of the offshore wind farm collection system. When the voltage level of the collection system is low, it will face the problems of limited redundancy of the number of wind turbines in the series and a more complex topology structure of the collection system, resulting in low planning efficiency or planning failure. For 66kV and 35kV Power collection systems, we improved the FCM algorithm and prim algorithm from the aspects of edge weight conditions and optimal path selection, and reduced and partitioned the dimension of the overall planning. By improving the algorithm structure and optimization method, we decomposed the optimization of the complex power collection system into the optimization problem of multiple partitions. This makes the improved algorithm can be applied to the topology optimization of offshore wind farm power collection system with different voltage levels in the trend of large-scale fan capacity. Moreover, an example was carried out to verify that the optimization algorithm has good universality, optimization, and planning efficiency.