For the intelligent management of micro-energy grid, the limitations of traditional scheduling optimization methods have begun to be highlighted, and computer technology has become a new generation of power system support means. This paper proposes a state estimation method based on smart grid measurement technology, and for the characteristics of micro-energy grid anomaly data, adopts the sampling value detection anti-anomaly data method based on amplitude comparison to estimate the grid dynamic process. Based on the grid state estimation dataset, a smart grid scheduling strategy based on cloud computing is constructed. The conditional value-at-risk of the penalty function of generation cost, grid cost, and motion cost is used as the objective function of the dispatch modeling, and the solution method for the optimization of the value-at-risk model is provided. Finally, an example has been built for this dispatch model to access a simulated micro-energy grid system that contains multiple nodes. The results show that the error between the voltage amplitude obtained from the dynamic state estimation of the micro-energy grid using the method of sampled-value detection against anomalous data and the real value is no more than 0.5%, and the data scheduling error rate of the cloud computing does not fluctuate significantly when dealing with a large number of scheduling data tasks, all of which are controlled to be below 0.0015. The scheduling strategy based on cloud computing has a certain degree of generalization when dealing with random and variable source and load scenarios, which can effectively improve the level of power grid business applications and complete the corresponding intelligent scheduling.