The application of electromagnetic signals in earthquake study has been applied by previous researchers through the monitoring of geomagnetic variations. The previous studies have revealed inconsistencies in the implementation of the diurnal variation ratio (DVR) method and the results were also found to be limited in specific events. This study sought to enhance the reliability of earthquake forecasting by implementing two different variants of the DVR method in investigating the magnetic responses prior to earthquakes (EQ). Global EQ events that occurred between 2000-2020 with magnitude above 5.0 were observed. The anomalies were detected as early as 60 days to 1 day prior to the EQ events for DVR using threshold value (Method 1), and 30 days to 15 days prior to the EQ events for DVR using the comparison with 1-year background geomagnetic data (Method 2). All geomagnetic N, E, and Z components showed anomalous behaviour during the quiet days but with temporal lags between the components. It can be concluded that Method 1 approach, yielded results with significantly more precursor presence than Method 2. The relationship of the geomagnetic variations with earthquake properties such as magnitude and focal depth showed higher rate of precursor presence in both the strong and mid-focus EQ. Future studies will be conducted to correlate geomagnetic variations with seismo-ionospheric response and physical ground movement prior to the events. The outcomes of this study will be able to provide insights of effective analysis for precursor study particularly in seismic hazard.