The purpose of this study is to propose a nearest neighbor search method for reverse geocoding on mobile tracking system. This method is based on some characteristic of mobile tracking system and database management system (DBMS). "Reserve Geocoding" is the process of finding place names, street addresses and other resources for a point location (latitude and longitude). We can find some of ready-to-use reverse geocoding API on the internet. Generally, it searches from the database that includes latitude, longitude, place name and so on, and that database is huge number of records. The place identified by addresses is not a point, but it is a complex polygon. A search method is to calculate whether the polygon contains a target coordinate, and is no simple algorithms and required high-performances computer system. A proposed method is do search by indexed coordinates (latitude and longitude), next do nearest neighbor search. The database is indexed by latitude and longitude, so it takes to high performance response. By using an indexed database and a proposed method is applicable to shortens the search time, and results of the experiment on the mobile tracking system showed that a method was a practical performance.