Abstract: This paper introduces the frequently used methods to analyze and process georesistivity observation data for the purpose of earthquake prediction, which include the improvements to some previous methods and the new methods since the ninth Five-Year Plan. In all, there are the eight methods that can be classified into four categories, that is, method of eliminating annual variation, dimensionless method, relative root-mean-square-error (RMSE) method and difference method. Some specific problems about the eight methods are discussed in detail, such as method principle, procedure of data processing, approaches of anomaly analyses, resolution capability for anomaly, indices for identifying anomaly, physical mechanism of anomaly as well as their shortcomings. The results show that: ① in general, effective data processing methods ought to be used to analyze and identify the "weak change" anomaly on the raw data curve; ② the principle of the eight methods are simple and clear, the physical mechanism on georesistivity anomalies based on the methods are clear or more clear, and the indices for identifying anomaly of each method are definite, relatively definite or qualitative, respectively. The qualitative anomalies are used only for reference in actual earthquake analysis; ③ the method of eliminating annual variation and dimensionless method is usually used to identify the anomalies in the medium-term and imminent periods before an earthquake, and the latter two kinds of methods are usually used to identify imminent anomalies; ④ the raw observation data from a station is true, so the anomaly got through a data processing method should be concordant with the "weak change" anomaly on a raw curve; ⑤ the anomaly arising in the time series of observation data, called as "data anomaly", does not equate with the precursory anomaly directly related to the preparation and occurrence processes of earthquake, and it is not expected that a noticeable earthquake will be certain to occur once any data anomaly appears on the time series of data.