This paper presents a language identification system based on GMMUBM algorithm, the system combines the border of Yunnan radio regulatory characteristics, depending on the frequency, receiving location, etc., using different identification sublibraries for the identification of crossborder signal can be better improvement of crossborder due to poor signal quality monitoring to lead to a simple algorithm using GMMUMB recognition rate not high. After Mans monitored for crossborder recognition experiment actual signal, identifying the effect of significantly improved, demonstrate the effectiveness of the system.