Using the Cubature Kalman filter (CKF) method to estimate the state of non-cooperative targets in space can mitigate the impact of Gaussian noise. However, in the complex environment of space, due to its propagation conditions, there are still a large number of non-Gaussian noise and sudden outliers that cause estimation bias. Considering the limited observation range and state estimation accuracy using a single chasing spacecraft, this paper proposes a multi-spacecraft observation scheme with data fusion. This scheme integrates observations from multiple spacecraft through an information filter and employs the CKF method based on a robust strategy to remove the impact of Gaussian noise, non-Gaussian noise, and sudden outliers in space, achieving stable estimates. Through a comparative study with the single-spacecraft strategy, it is demonstrated that employing multiple spacecraft in this scheme can effectively enhance the estimation performance and stability. Furthermore, by comparing different estimation methods under multi-spacecraft scenarios, the effectiveness of the robust strategy selected for this scheme is validated. To conclude this paper, we provide our expectations for further work on target non-cooperative spacecraft perception.