Double-sided solar greenhouses are recognized as energy-efficient agricultural facilities that significantly enhance land utilization and thermal performance through their unique double-sided lighting design, thereby promoting crop growth. However, challenges persist regarding insufficient heat storage capacity and suboptimal thermal environments within the shaded shed during the winter and spring seasons. To fully exploit the advantages of this greenhouse type, this study proposes a structural optimization methodology utilizing Computational Fluid Dynamics simulation. A CFD model was developed and validated against experimental data to ensure accuracy. Subsequently, the influence of key parameters, including roof geometry and wall thickness, on the internal photothermal environment was systematically analyzed. The results demonstrate that the 370 mm thick wall configuration achieves a daily peak temperature approximately 2°C lower than the 240 mm wall, indicating a more uniform spatial distribution, while exhibiting a nighttime temperature increase of up to 2.5°C, thereby confirming superior thermal insulation properties. Furthermore, the presence of a rear roof structure is critical for nighttime heat retention, maintaining a minimum temperature of approximately 5°C compared to 2°C in greenhouses lacking this feature, with a maximum temperature difference of 4.2°C, effectively optimizing temperature uniformity. Based on these findings, this research provides a robust theoretical foundation and technical support for the structural optimization of double-sided solar greenhouses and the advancement of facility agriculture.