A new nonlinear robust filter is proposed to deal with the outlier of GPS/INS integrated navigation system. The influence of different design parameters for the H_∞  cubature Kalman filter is analyzed. The design parameter is smaller, and the robustness of the filter is stronger. But the design parameter is easily out of step with the Riccati inequation and the filter is easy to diverge. In this respect, the singular value decomposition algorithm is used to take the place of Cholesky decomposition in H_∞  cubature Kalman filter. On the wider conditions for the design parameter, the new filter is more robust. The actual GPS/INS integrated navigation test indicates the correctness and effectiveness of the proposed filter algorithm.