Abstract: For multi-Global Navigation Satellite System constellations, satellite selection can be considered as a constrained multi-objective optimization problem to optimize the Geometric Dilution of Precision (GDOP) and number of selected satellites simultaneously, as a result reduce computation of the receiver and improve positioning accuracy. Here we describe a method based on the NSGA-Ⅱ algorithm, which utilizes the sequential characteristic of satellite selection to generate the initial population, as well as improves constraint handling, appropriate genetic operatorand utility function to achieve fast satellite selection.The simulation results show that the proposed method has high reliability and real-time performance and does not rely on the geometric distribution of satellites, making it suitable for complex environments with obstacles and shelters.