With the development of the Global Navigation Satellite System (GNSS),the number of visible satellites increases greatly.In order to lower the computational complexity of the receiver,a satellite selection algorithm is essential for choosing a subset of the visible satellites.This paper analyses the Global Dilution of Position (GDOP) of a joint distribution,which contains the N optimal distributions of four satellites and the k (k≤3) satellites in an optimal distribution of four satellites.According to the comparison of the GDOP between the joint distribution and the optimal distribution of 4N+k satellites,a four-based satellite selection algorithm is proposed,which has low computational cost and is close to the optimal.The proposed algorithm can select any number of satellites more than four through the iteration and portion of the classic satellite selection algorithm of four satellites,such as maximum-volume algorithm and four-step satellite selection method.Simulations indicate that the proposed method requires less computational cost and achieves lower GDOP than the quasi-optimal algorithm.