Because of the special environment of the underground railway train,and the different operating condition(maneuvering or nonmaneuvering), with traditional Kalman Filter,it is difficult to react filtering error problem in the mine realtime positioning system. proposes the data processing method with interaction multiple model Kalman (IMMKalman) Filter.the way adopts IMMKalman to filter processing position of underground ailway train estimated by wifi and microinertial unit combination location,The simulation results show that the IMMKalman can effectively reduce the realtime positioning error，so can enhance realtime performance and positioning accuracy.