This paper discusses the design and research of smart grid based on artificial intelligence technology, which realizes automatic control and optimization of power system by applying particle swarm optimization algorithm and multi-objective optimization theory, in order to improve the operation efficiency and stability of power grid. In view of the current grid distribution chaos, data error consumption and other problems. Using particle swarm optimization algorithm and multi-objective optimization theory to design and research smart grid principles and methods, particle swarm optimization algorithm can find the best solution in the search space, and through the multi-objective optimization theory to balance different target requirements, so as to achieve the automatic control of power system. Deep neural networks are trained and utilized with large amounts of data for prediction and classification purposes. By using a variety of means, such as reinforcement learning, we get a diagnosis scheme of intelligent power system, and build a relatively perfect intelligent power management software. The results show that the new smart grid design system is more economical and simple than the traditional grid distribution. Better solutions for saving electricity.