Tracking enough visible GNSS satellites is becoming more and more significant, because of the growing demand of higher positioning precision. More and more satellites are launched, which is beneficial to positioning and causing huger tracking burden in the meantime. In this paper, we have integrated the genetic algorithm into multi-constellation GNSS satellites selection and analyzed its performance. Our efficient genetic algorithm of multi-constellation satellites (GPS, BeiDou, Galileo and Glonass) selection(EGAMS), whose parameters are determined by grid search method, will give the optimal satellite's combination candidates. The simulation results between EGAMS and traditional optimal satellite selection algorithm (TOSSA) indicate that EGAMS has a significant improvement on the computational complexity with the same accuracy according to WGDOP. In the meantime, the time consumption of EGAMS dramatically decreases 75% compared with TOSSA.