This research addresses the enhancement of navigation accuracy and robustness in rocket formation landings, which can improve the reliability of multi-booster heavy-lift reusable rocket launch and recovery missions, as well as the formation landing of probes in deep space exploration. Utilizing a sophisticated data fusion algorithm that integrates a dual quaternion dynamic model with an Unscented Kalman Filter (UKF), this study harnesses data from ground radar, Inertial Measurement Units (IMU), and six-degree-of-freedom optical tracking modules. Extensive simulation tests have demonstrated that the data fusion method leads to a slight increase in pose measurement accuracy. Additionally, a FDI mechanism enhances the system’s fault tolerance, enabling detection and isolation of two concurrent sensor failures in a trio of rockets. This finding elevates navigation precision and sensor failure management, potentially benefiting future space missions.. ...