Electric vehicles (EVs) are getting popular as one of the effective solutions for increased energy efficiency in commercial systems. This paper proposes an improved algorithm for commercial peak-load management using EVs, battery-energy-storage systems, and photovoltaic units. It uses the bidirectional vehicle-to-grid technique to utilize the energy from EVs in a parking lot. The proposed system has been tested in a real power distribution network in realistic load and weather conditions. The financial benefit of the system is also investigated, and it is found that the industrial peak load can be reduced by 50%, and the energy cost can be reduced by up to 27.3%. It also enhances the load factor by 9%. The performance of the proposed control algorithm is compared with that of an artificial-neural-network-based technique and tested in a laboratory prototype. From simulated and experimental results, it is found that the proposed approach provides substantial savings, while reducing the peak demand of the existing grids.