This study proposes a novel angle-based partition design method to optimize the collector system of offshore wind farms. It highlights the importance of improving the reliability of the collector system while reducing overall costs. By integrating the minimum spanning tree algorithm with angle-based partition theory, this method minimizes subsea cable lengths and reduces failure risks under constraints such as the number of feeders, subsea cable current capacity, balanced feeder power, and avoidance of cable crossings, thereby enhancing system efficiency and reliability. A case study on a recent 1-GW offshore wind farm in Guangdong, China, was used to validate the effectiveness of the proposed method. A multi-criteria decision analysis approach was used to evaluate the proposed method. The evaluation results showed that the design scheme generated by the angle-based partition method performed the best. This design scheme demonstrated a more balanced distribution of turbines per feeder, with virtually no increase in subsea cable length compared to traditional methods such as Improved Prim and Angle-based K-means Clustering.