Modeling the neutral density in the upper atmosphere is a major challenge as it involves a comprehensive understanding of the complex coupling between the thermosphere and its environment. In addition, this system is strongly driven by solar activity, which is hardly predictable. This complexity limits the ability of thermospheric models to accurately estimate thermospheric characteristics, such as the neutral density. With the rapidly increasing number of satellites in orbit and the recent progress in tracking their trajectories, new techniques are being implemented to derive density estimates based on trajectory perturbations. However, most of these methods require the integration of complex and costly orbit determination algorithms, which greatly limit their applicability at a public level. The approach presented here provides density estimates along a spacecraft trajectory and only requires the implementation of an orbit propagator. More specifically, it corrects density predictions made by the NRLMSIS00e thermosphere model by optimizing the orbit fitting of modeled trajectories with observations under different thermospheric density conditions. The algorithm is applied to the Cyclone Global Navigation Satellite System and Swarm-B spacecraft trajectories during a period of varying geomagnetic activity. The study puts in evidence a bias in NRLMSIS00e estimates during moderate solar activity, corrected by the algorithm. In addition, the approach provides a better characterization of the rapid density enhancement resulting from a geomagnetic storm. Overall, the correlation with the Level 2 data is improved by 15% and the normalized root mean square error decreased by a factor ~2.3, compared to the estimations with NRLMSIS00e.