Optimizing the pore structure of the cathode catalyst layer of proton exchange membrane fuel cells (PEMFC) is crucial for improving oxygen transport and overall cell performance. In this paper, the simulated annealing (SA) algorithm was improved to iteratively optimize the catalyst layer reconstruction model, improve the algorithm's computational efficiency, and construct a model with physical properties close to the pore structure of the actual catalyst layer. The phase exchange process during the computation is performed using an improved Different Phase Neighbors (DPN) algorithm, which prioritizes the nodes with significant phase differences for phase exchange and improves the efficiency of generating candidate solutions. The Hierarchical Annealing Strategy accelerated convergence, reducing computational complexity without compromising accuracy. Comparative analyses of pore size distributions and performance evaluations between the reconstructed models and the actual CL structures confirmed the model's effectiveness and accuracy.