Abstract: To improve the design methodology for composite cooling structures of turbine blades, this study investigates the flow network method applied to the design, analysis, and optimization of cooling configurations. Building upon traditional one-dimensional flow network calculation methods, a coupled design and computation platform was developed by integrating internal and external blade cooling considerations. Taking the NASA GE-E³ first-stage rotor blade as a benchmark, a preliminary cooling scheme was designed and subsequently validated against three-dimensional conjugate heat transfer simulations. Based on the initial design, the blade cooling structure was redesigned, and the ribbed channel flow network was optimized using the NSGA-Ⅲ algorithm. Results show that the relative error between the one-dimensional flow network predictions and the three-dimensional conjugate heat transfer results is below 2%, confirming the computational accuracy of the proposed method. The network optimization was completed in approximately 25 h,and the blade temperature after modification and optimization is obviously reduced. The proposed method demonstrates high computational fidelity and offers practical guidance for designing composite cooling structures in turbine blades.