Abstract: A novel discrete artificial fish school algorithm (DAFSA) was proposed,in which a new strategy is designed about fish moving behavior,improved the optimization ability of AFSA in combinatorial optimization problems and Partial transmit sequence (PTS) algorithm.By replacing the corresponding element of artificial fish in the current position with the optimal position vector to realize the position update,which made the search trajectories converge to the best phase factor quickly,and then get the lowest peak-to-average power ratio (PAPR).The simulation show that DAFSA-PTS have the nearest PAPR with conventional PTS,when the number of sub-block is12,has a differential of 0.4dB,and the complexity is reduced about 85.35%;when the complexity are same,DAFSA-PTS performs better than PSO-PTS,and the precision increases about 0.2dB.