The bakingBaking process critically determines the physicochemical propertiesProperties of carbon electrodesCarbon electrode used in aluminum productionAluminum production. While previous studies have correlated bakingBaking parameters with electrode performance, the formation mechanisms of porous structures remain poorly understood, limiting further optimizationOptimization. This study employs X-ray Computed TomographyComputed tomography (CT) and Digital Volume Correlation (DVC) to establish a mesoscale model of carbonCarbon cathode blocks and propose an intelligent optimizationIntelligent optimization model for low-defect formulations. Through 3D reconstruction, thermal deformation analysis, and pore classificationClassification, we quantitatively analyzed dynamic structural evolution during thermal treatment. Results show most pores form below 410 °C, with predominant many-to-many connectivity between secondary and primary pores from 240 to 410 °C. Based on fractal dimensions from CT analysis combined with densest packing theory and particle swarm optimizationOptimization, we developed a particle size distributionParticle size distribution model for low-crack blocks. The optimized formulation achieved 43.09% reduction in air permeability, 4.03% increase in density, 14.39% improvement in thermal conductivity, and 25.34% decrease in electrical resistivity.