The invention provides a data cleaning method and system for a GIS partial discharge big data system, and the method comprises the steps: building the GIS partial discharge big data system, collectingdata, formulating a unified data format, developing a conversion module, and carrying out the format normalization conversion of a heterogeneous data source; then, processing abnormal data through exploratory analysis, establishing corresponding cleaning rules for different data types forming a data cleaning rule base , arranging the data into a data set which is uniform and accurate in format, and continuously optimizing the data cleaning rule base; and finally, performing mining analysis on the data. According to the invention, format conversion and data cleaning can be carried out on the partial discharge detection data, analysis and mining are carried out by using an advanced algorithm, and the diagnosis accuracy of the partial discharge detection data is improved.