Abstract: In order to maximize the uplink achievable sum rate of the multi-cell and multi-user massive multiple-input multiple-output (MIMO) system, a pilot allocation algorithm based on matched game (PA-MG) between users and pilots is proposed. Firstly, on the user side, a preference list of pilots is generated according to the user utility function, and an application is made to the pilots with the best arrangement for all users. Then, on the base station side, the preference list of the requesting users is generated according to the pilot utility function, and the pilots are assigned to the users with the best arrangement in turn until all users are assigned with the pilot. Simulation results show that the PA-MG algorithm supports larger cellular and less time complexity than the pilot allocation based on potential game (PG-PA) algorithm. Compared with the weighted graph coloring based pilot decontamination (WGC-PD) algorithm, the PA-MG algorithm can achieve greater uplink achievable sum rate, making the distribution of signal to interference plus noise ratio (SINR) of each user be more uniform, and the robustness to shadow fading is stronger.