Earthquake prediction is a long-term concerned but unsolved scientific challenge in the global seismological community. Due to the complex seismogenic process and the inaccessibility of underground structures for direct observation, traditional geophysical methods still have considerable uncertainty in the accurate prediction of earthquake occurrence time, location and magnitude. In recent years, geomagnetic methods have gradually become one of the important directions in short-term earthquake prediction research, owing to their advantages such as high sensitivity to changes in underground electrical structures, mature observation techniques and strong data continuity. In this paper, the daily ratio method of the vertical component daily variation amplitude of the geomagnetic field is used to study the relationship between the daily ratio of the geomagnetic field and the earthquake with M_S≥5.0 in Xinjiang from January 1, 2015 to December 31, 2025. The results indicate a strong correlation between the daily ratio of the geomagnetic field and M_S≥5.0 earthquakes. Within months after the daily ratio of the geomagnetic field exceeds an abnormal threshold line, the probability of an earthquake with M_S≥5.0 occurring in the corresponding area is high. The following conclusions are drawn.

1） Based on the minute-sampled data of the geomagnetic vertical component from the National Geomagnetic Network Center, this study investigates 43 earthquakes of M_S≥5.0 in Xinjiang, China from January 1, 2015 to December 31, 2025. A total of 32 groups of geomagnetic daily ratio anomalies are counted, and 19 groups are followed by earthquakes. The anomaly correspondence rate is 59.4%, the false alarm rate is 40.6%, the earthquake missing rate is 44.1%, 21 earthquakes are successfully predicted, and the earthquake prediction accuracy is 48.8%. The dominant corresponding period is 2 months. Among them, only one geomagnetic daily ratio anomaly appeared before seven earthquakes, while two or more anomalies occurred before 13 earthquakes. Multiple anomalies are common before earthquakes, and the false alarm rate of grouped anomalies is much lower than that of single anomaly. Due to the uniqueness of geotectonic and active tectonic conditions in Xinjiang, the geomagnetic daily ratio anomalies before strong earthquakes also show regional characteristics. From January 1, 2015 to December 31, 2025, there were 25 days with geomagnetic storms characterized by the daily mean index D_st＞−70 nT in the global geomagnetic environment, and consecutive storm days were counted as one event （D_st indices are obtained from https://wdc.kugi.kyoto-u.ac.jp/）. Among them, six storms triggered over-threshold geomagnetic daily ratio anomalies. Only three of these storm-induced anomalies were followed by three earthquakes with M_S5.0−5.9 within three months, and no earthquakes with M_S≥6.0 occurred. The remaining three storm-induced anomalies were not followed by any earthquake of M_S≥5.0. Statistics show that there is no one-to-one correspondence between moderate-strong geomagnetic storms and geomagnetic daily ratio anomalies, and most anomalies occur on geomagnetically quiet days. The monthly frequency of geomagnetic daily ratio anomalies is generally higher in winter and summer and lower in spring and autumn, without an obvious dominant distribution.

2） This paper also studies three strong earthquakes with M_S≥6.0: the M_S6.1 Shaya earthquake on January 30, 2023, the M_S7.1 Wushi earthquake on January 23, 2024, and the M_S6.0 Akqi earthquake on December 4, 2025. It is found that the frequency and spatiotemporal evolution of geomagnetic daily ratio anomalies before these three earthquakes show certain similarity and repeatability. That is, two or more over-threshold geomagnetic daily ratio anomalies appeared within three months before these three earthquakes. This feature may indicate multiple micro-fractures in deep structures shortly before the mainshock, which can be regarded as a precursor of the sub-instability state before earthquakes. Before the mainshocks, geomagnetic daily ratio anomalies migrated from distant areas to the epicenter, with anomaly areas shrinking from large to small or moving toward the locked seismogenic zone. Epicenters are mainly located at the intersections, concentrated zones and high-curvature sections of anomaly threshold lines, which is helpful for predicting the location of future earthquakes.

3） By superimposing multiple daily geomagnetic ratio anomalies threshold lines on the same base map, an abnormal intersection area or a concentration area can be formed.

Earthquakes frequently occur in these areas where daily geomagnetic ratio threshold lines are concentrated or intersect. Combined with the distribution of active tectonic faults, historical seismicity, multi-parameter anomaly coupling, and the long-term and short-term predictions for the potential strong earthquake risk zones and annual seismic hazard zones in Xinjiang, comprehensive judgment on seismically risky areas can effectively reduce the errors in location, time and magnitude predicted by the geomagnetic daily ratio method. This provides a reference for determining the location and time of subsequent strong earthquakes, and is of practical significance for the application of earthquake prediction techniques.