With the continuous improvement of the intelligent level of the power system, the data system generated in the power grid is also becoming larger and larger, and the quality of the data will directly affect the operational analysis and planning decisions of the power system. Therefore, this paper proposes a power grid time-sequence data quality maintenance system based on data mining technology to screen out unqualified data to ensure the correctness and reliability of the acquired data. Meanwhile, it can identify problems in the data and help analyze the cause of the problem. Firstly, this paper analyzes the power data and transmission process, and points out the possible problems. The data of different regions have their own different characteristics. In order to improve the detection speed, this paper first makes decision analysis on historical data samples based on decision tree algorithm. This paper takes the data training set of a certain area as an example to analyze the power data detection process in this area, and obtain the detection sequence suitable for the area. Then, for the problem that the data rationality is difficult to detect, this paper adopts the cluster-based outlier detection method to filter the data that does not meet the operational requirements and try to analyze the cause of the problem data. Finally, the effectiveness and reliability of the time-sequence data quality maintenance process proposed in this paper is illustrated by an example.