Aiming at the cloud computing cluster environment, large-scale data processing of smart grids is under great pressure. Task scheduling and rapid processing of large-scale data distribution are urgent problems to be solved.Based on cloud computing and big data processing technology, set the multi-vibration priority algorithm and node selection algorithm for task scheduling and publishing,which can process and distribute important data in a timely and reliable manner.This algorithm is compared with the classical algorithm through simulation to verify the accuracy and effectiveness of the algorithm.The results show that this method can effectively distribute node load and has obvious advantages in task completion rate and completion time.This research provides reference and reference for the research of China"s electric power big data information platform.