The invention discloses a missile-borne deep integrated ARCKF (Adaptive Robust Capacity Kalman Filtering) method under a strong maneuvering condition. The missile-borne deep integrated ARCKF method comprises the following steps: generating a ballistic trajectory and corresponding IMU (Inertial Measurement Unit) data of a simulated strong maneuvering missile through a trajectory generator; inputting the generated ballistic trajectory into a satellite signal simulator, and generating a GNSS (Global Navigation Satellite System) intermediate-frequency signal; then inputting the generated GNSS intermediate-frequency signal into a software receiver and carrying out inertial navigation calculation on the IMU data; establishing a state equation and an observation equation of a GNNS/SINS (Strapdown Inertial Navigation System) deep integrated navigation system under a launch inertial coordinate system; simultaneously integrating a robust M-estimation algorithm in a robust estimation theory and adaptive factors to a CFK (Capacity Kalman Filtering) algorithm, forming an ARCKF algorithm, and carrying out filtering correction on a system state. The missile-borne deep integrated ARCKF method disclosed by the invention can be used for GNSS/SINS deep integrated navigation of a carrier under strong maneuvering and high dynamic conditions, and the anti-jamming capability of navigation and the navigation accuracy are effectively increased.