In this project, a unified consistency based Gaussian mixture (CbGM) tracking framework is proposed for the detection, tracking and identification of space objects using measurements from disparate sensor networks. Major components of the proposed tracking framework are as follows. First a Consistency based Gaussian Mixture approach which is more efficient than other existing Adaptive Gaussian Mixture approaches is proposed for the accurate propagation and update of space tracks. Second, a novel Sparse-grid Quadrature Filter (SGQF) which offers improved accuracy over the UKF with moderate increase of complexity will be used in the CbGM framework for the tracking of space objects. Third a novel Rollout Policy based Multi-step Look-ahead Data Association (RO-MSLHDA) approach is proposed for accurate multi-target tracking (MTT) in complex space environment. By taking advantage of the highly deterministic nature of orbital motion and exploring information from future observation frames, the RO-MSLHDA is developed to achieve tracking accuracy that is close to that of the Multiple Hypothesis Tracking (MHT) with much lower complexity. The impact of dynamic model accuracy on data association performance and effective track initialization using existing Initial Orbit Determination (IOD) algorithms with the RO-MSLHDA framework will also be investigated.