This paper describes a ground-based orbit determination algorithm for the remote sensing satellite BILSAT. A near-real time quality control of orbit data in the ground station is of vital importance to detect outliers or to bridge data gaps where position information is not available. These might result both from possible temporary mal-functioning and intermittently on/off switching of the onboard GPS receiver. A dynamic approach is preferred for the orbit determination because it includes not only outlier's detection or bridging of data gaps without GPS observations but also rigorous modeling of relevant forces with a better model accuracy, reliable velocity information of the space craft and direct estimation of other important parameters like atmospheric drag or radiation pressure coefficients. Such an algorithm provides continuous improved orbit information on the basis of force models and space-borne GPS position fixes as observations. In context with the dynamic model a best fitting orbit for BILSAT is then achieved by using a univariate Gauss-Markoff model as the estimation procedure so that improved orbit information in terms of Cartesian coordinates with respect to WGS 84 can be provided.