This paper aims to propose a set of lost-in-space star identification algorithm that works effectively for small satellites. Several algorithms are investigated for each phase of star identification: feature extraction and catalogue searching. For feature extraction, it is shown that the algorithm using extended images and that using combined images work best, and for catalogue searching, the Group Catalogue is shown to work best. Therefore we propose an algorithm combining these three as the best one. Simulation results also show the effectiveness of this algorithm in its robustness to various errors.