We systematically investigated precursory seismic patterns using the pattern informatics (PI) method and suggest an operable procedure for making PI maps for all seasons, in the context of earthquake forecasting. We examined the PI patterns before several inland earthquakes with magnitudes larger than 6, which occurred between 2001 and 2010 in Taiwan. We fixed a cutoff magnitude and a change interval, which is the time span used to calculate the seismicity change. Our results show that locations with high PI anomalies are typically associated with large earthquakes when the cutoff magnitude is 3.2 and the change interval is 4 years. Therefore, the PI method can be utilized as a routine forecasting tool with regular updates, such performing the PI calculation every season. We also conducted random tests, the results of which indicate a significant difference between large events and random, hypothetical events.