With the intelligent space technology, a composed global localization system based on ZigBee fingerprint and improved particle filter is proposed for solving substation robot global localization problem in which the initial pose of robot is unknown and the robot pose is determined according to sensor information under large scale outdoor environment, which is of great significance for the construction of smart grid. In the composed localization method, initial location is ?rstly realized using the ZigBee ?ngerprint localization based on BP network under the support of existing wireless network in substation, and then on the basis of initial location, particle swarm optimized particle filter is used to locate accurately. The experiment results show the proposed system can realize zero configuration and loose coupling between substation robot and intelligent space, improve global positioning accuracy and the performance of the algorithm, and reduce the iteration time.