Abstract: Large scale air conditioning load in summer intensifies the peak and valley difference of power grid, which has a serious impact on its economic operation. Therefore, it is significant to seek effective measures to reduce the load of air conditioning. In this paper, the energy management of solar building is studied considering solar radiation uncertainty. In order to make full use of solar energy resources, the building is configured with photovoltaic/photo-thermal (PT) integration system, whose PT part is aimed to collect thermal energy for air conditioning system to reduce electricity demand and photoelectricity part is used to supply electricity stored in battery for night load power. As to issue of uncertainty, the error probability model is introduced, in which Monte Carlo simulation method is applied to calculate electricity costs. In addition, according to different characteristics of controllable load, the operation periods of electrical equipment in building are optimized to reduce electricity cost under real-time pricing environment. In case analysis, the electricity economy of solar assisted system with PV/PT devices and traditional system is compared and analyzed.