This work introduces a near real-time dynamic orbit determination program for small satellites. A quality check of orbit data on ground is important for satellites with GPS positioning that is only available at discrete time epochs because of the limited power-supply onboard, so that space-borne GPS receivers are switched on/off intermittently, or because of possible malfunctioning of GPS receivers. The method of parameter estimation with space-borne GPS position fixes as observations is based on an extended Gauss-Markoff model which corresponds to a Bayesian least-square estimation with informative priors for the parameter vector. Such an approach in context with a dynamic model determines a best-fitting satellite orbit and allows detecting outliers or bridging of data gaps where GPS measurements are missing. It can also estimate other important parameters such as the satellite's drag or radiation pressure coefficients. Observations and parameters are dynamically managed by a database management system. The attached program is written in Visual Basic 5, C/C++ and runs on MS Windows 95/NT or later. (c) 2004 Elsevier Ltd. All rights reserved.