The invention relates to the technical field of power data cleaning, in particular to a power missing data filling method based on a hybrid strategy. The method comprises the steps that S1, an improved k-Means clustering algorithm is adopted to perform clustering on a dataset containing missing data; S2, an RBF neural network is improved and constructed according to the clustering result; and S3,the RBF neural network is trained, and filling checking is performed on the missing data. Through the method, clustering of the dataset with missing properties is well realized, and the RBF neural network is designed in combination with the clustering result to perform prediction filling on missing values. Besides, the filling precision of the missing data is improved, implementation is easy and convenient, calculation expenditure is appropriate, and the method has high practical value against the problem that a large amount of data generated in the operation and maintenance process of a powersystem is missed and damaged due to the influences of physical factors, software factors and other factors.