In order to improve the navigation precision of strapdown inertial navigation systems/celestial navigation systems (SINS/CNS) integrated navigation system in high dynamic conditions, a SINS/CNS integrated navigation algorithm which based on the additive dual quaternion was proposed. The rotation and translation of carrier were unified by this algorithm, and the rotation vector was used to update dual quaternion and compensate coning error and sculling error at the same time. And derived error model based on the additive dual quaternion and error equations of navigation parameters. Then the constant errors of gyro and accelerometer were extended to be state variables, at the same time random error was used as the system input noise, and gyro drift was corrected by output parameters of star sensors. Finally, Kalman filter was used to estimate the state variables. The simulation results showed that the navigation precision of SINS/CNS integrated navigation algorithm based on the additive dual quaternion is increased by two times more than traditional algorithm in high dynamic conditions.