Windows with controlled transparency are of increasing interest due to their potential energy saving and variable architectural view. These “smart windows”, are a promising approach to reduce the energy consumption of buildings, which contributes up to 40% of the world’s energy usage. There are several types of them, for example the ones which are voltage controlled such as the liquid crystalline based ones or the electrochromic materials usually control the transparency of the whole solar spectrum. However, thermochromic materials such as VO2 based, they block a large fraction of the sunlight on hot days, while transmitting solar energy in cold weather. Thus through temperature or voltage responsive solar energy modulation, smart windows are a key component of green buildings that have been extensively studied in the recent years. During the last few years we have been studying several concepts to improve smart window functionality both using thermochromic VO2 based nanostructures and recently based on liquid crystal line nanoporous micro-structure. In this talk we shall review our latest results along the following lines:Liquid crystalline new composite metamaterial made of nanoporous microparticles infiltrated with liquid crystal and controlled by temperature and voltage. VO2 thermochromic nano-grid films with high luminous transmittance while maintaining the solar modulation ability. The perforated VO2-based films employ orderly-patterned nano-holes, which can favorably transmit visible light dramatically but retain large solar modulation ability. SiO2/VO2 2D photonic crystals engineered for structural color tuning of reflected and transmitted light, while maintaining high solar modulation. Au/VO2 half sphere core-shell structure as a case study of the temperature-dependent plasmonic effects and the thermochromic response for smart window applicaton. By varying the Au core size, we were able to tune the transition temperature of VO2 shell. The analysis of the induced thermal strain at the Au/ VO2 interface shows that this could be the dominant reason for transition temperature reduction, contrary to the widespread view that plasmonic effects play the main role. Design rules for optimum performance of smart windows will be presented including a new approach based on deep learning.