Strapdown Inertial Navigation System (SINS) and odometer (ODO) are both autonomic devices. The SINS/ODO integrated navigation is a useful method in land positioning and orientation applications. Especially for the global navigation system signal denial (GNSS-denial) battlefield, the optical gyroscope SINS/ODO integrated navigation is fully capable of long-distance high-precision navigation due to its autonomy. However, due to the uneven distribution of the earth mass, there exists a discrepancy between the the gravity vector calculated by the traditional model and the real gravity vector, that is, gravity disturbance and anomaly, which leads to positioning error in high-precision SINS/ODO integrated navigation. Therefore, this paper firstly analyses the influence of gravity disturbance on the state-transform extended Kalman filter (ST-EKF) based integrated navigation system, and gives the local gravity disturbance calculation method using the EGM2008 ultra-high order spherical harmonic model. A deterministic input compensation method using gravity disturbance is proposed, and the effectiveness of the gravity disturbance compensation approach is verified by two groups of long-distance high-precision SINS/ODO integrated land vehicle experiments.