Analysis of spoken and search engine queries show that people often look for maps of a certain location that are at a certain time or show a subject theme. Their search may be hindered by the fact that maps in both physical and digital collections are often organized by region, and maps in non-geographical publications are essentially invisible to search because they are rarely indexed. This research offers a way to resolve the problem by writing algorithms that automatically classify maps by time and theme as well as region, demonstrating these techniques in a web-based prototype system. The classification is intelligent because it is aided by knowledge resources. These automatic classification were evaluated by comparing system-classified maps to manual classification of the same maps, with results showing automatic correspondence with the manual at 75% for region, 69% for time, and up to 93% for theme (when plausible classification categories are counted). The discussion will appeal to those who study information retrieval as well as to those who just like maps.