Since cross-fault means have been determined for nearly 40 years, the Sichuan area has accumulated rich observational data. The Sichuan region is an earthquake-prone area and, in recent years, the Sichuan Province experienced the Wenchuan and Lushan earthquakes, which were greater than magnitude 7. These huge earthquakes resulted in a great human casualty toll and property losses. As such, conducting earthquake prediction research has great significance. There have been many studies in this field and some important achievements. In this paper, we provide an overview of Sichuan earthquakes above 7.0 before which site anomalies have been observed. We also summarize the methods with which anomaly conditions were verified prior to the two large local earthquakes. In a recent analysis, an unusual approach involving cross-fault monitoring data was used, in which the authors found the faulting rate to reflect raw-data anomalies that eventually became abnormalities. On this basis, we introduce a wavelet analysis method for determining the presence of abnormality prior to an earthquake. We analyzed two trends in wavelet decomposition entries and found earthquake occurrence to correspond with abnormal wavelet decomposition. Finally, based on our analysis of raw data and wavelet decomposition, we propose the analysis of cross-fault earthquake precursory data as a reference for future research of earthquake abnormalities. In this paper, we statistically analyze horizontal and vertical anomalies greater than those of the M_S7.0 earthquake in Sichuan Province and use wavelet analysis to analyze long-term observation data. We conclude that the cross-fault monitoring data shows good ability to reflect earthquake occurrence. At present, there are many methods for monitoring earthquakes and in recent years, earthquakes have frequently occurred. Our proposed comprehensive method for conducting retrospective earthquake research provides a very useful prediction tool.