This paper describes an attitude determination experiment to be implemented on-board the ESEO satellite: An ESA sponsored LEO microsatellite to be launched in 2015 and currently under critical design review. The satellite attitude sensing suite include magnetometers (MGM), an Earth sensor (EVS), and a system of Sun sensors (SVS), each charactrized by a subdegree accuracy. The experiment will consists in running four algorithms that belong to the realms of the q-method and Kalman filtering, and implement varying sizes of state vectors for quaternion, angular velocity, and sensors' biases estimation. The sensors are activated by pairs: MGM-SVS in daylight and MGM-EVS at night. Sample runs comparatively illustrate the achievable performances - About 0.05 degree for all filters - with different dispersions and convergence times. Standard deviations plots clearly illustrate the changing observability of the parameters. A hardware testing campaign was conducted by running the algorithms on a microprocessor that is identical to the flight model. Tests successfully showed matching to numerical precision between the microprocessor-run C-code and the computer-run Matlab code. Other tests were performed to verify the correct management of telelcommand/telemetry/housekeeping data for initialization and activation of the various filters. Additional tests successfully verified the requirement on the maximum run time per cycle - 70 ms - Allocated by the on-board computer for the attitude determination experiment.