Abstract: To solve the invalidation of integrated navigation algorithm due to low updating frequency of global positioning system (GPS) in the low-cost integrated navigation system, a new integrated navigation algorithm, based on modeling of the inertial devices and kinematic modeling of the carrier, was presented. In this algorithm, the nonlinear system model was established. The attitude, velocity and position of the carrier were chosen to be the system state, and the output of the GPS, gauss meter, and altimeter were determined as the system measurement. The Kalman filter was adopted for data fusion after linearization of the system model. The static test and field test were implemented to validate whether the algorithm was useful. The test results show that the algorithm can effectively ensure that the error mean of attitude is smaller than 0.12 degree, the error mean of velocity is not larger than 0.03 meters per second, and the error mean of position is not larger than 3.94 meters, which is satisfied the requirement of the primary trainer.