Compared with traditional fossil energy, new energy has the advantages of large reserves, great development potential, and no pollution. For the current rapidly growing electricity demand, it is of great significance to fully develop and utilize new energy. In this context, the necessity of new energy development and the difficulty of new energy utilization drive the rapid development of distributed energy(DE) optimization scheduling methods. Since the efficient operation of DE optimal scheduling relies on advanced communication technology, this paper implements DE scheduling in the cloud environment. In this paper, the objective optimization scheduling simulation experiment of the economic cost and environmental protection benefit of the DE system is carried out, and the DE system is optimized by three learning algorithms including Bayesian, decision tree and improved decision tree, and the DE economic scheduling problem is solved , and compared with the other two algorithms, the effect of improving the decision tree algorithm to reduce the economic cost is more obvious.