This study was carried out to optimize a computational model of a new underground passive solar greenhouse to improve thermal performance, storage, and saving of heat solar energy. Optimized and conventional passive solar greenhouse were compared in regards of indoor air temperature, irradiation, and energy demand. Six different materials were used in the conventional model. In addition, TRNSYS software was employed to determine heat demand and irradiation in the greenhouse. The results showed that the annual total heating requirement in the optimized model was 30% lower than a conventional passive solar system. In addition, the resulting average air temperature in the optimized model ranged from 􀀀4 to 33.1 C in the four days of cloud, snow, and sun. The average air temperature in the conventional passive solar greenhouse ranged from 􀀀8.4 to 24.7 C. The maximum monthly heating requirement was 796 MJ/m2 for the Wtype87 model (100-mm lightweight concrete block) and the minimum value was 190 MJ/m2 for the Wtype45 model (50-mm insulation with 200-mm clay tile) in a conventional passive solar greenhouse while the monthly heating requirement estimated 126 MJ/m2 for the optimized greenhouse model. The predictability of the TRNSYS model was calculated with a coefficient of determination (R2) of 95.95%.