To overcome problem of precision decreasing of conventional cubature Kalman filtering（ CKF） when BDS/INS integrated navigation system model is not certain

combine CKF with singular value decomposition（ SVD）

and introducing strong tracking filtering（ STF） theory

an improved strong tracking SVD-CKF algorithm is proposed. To improve stability of numerical calculation

SVD is used instead of the traditional Cholesky decomposition of standard CKF; fading factors are introduced in order to modify covariance of predicted error online based on strong tracking filtering（ STF） theory framework

while system model is uncertain or system status is mutating

the proposed strong tracking SVD-CKF can improve robustness of system. Simulation results show the effectiveness of the presented algorithm.