In order to further improve the performance of particle swarm optimization in discrete optimization problems, aiming at the shortcomings of stickiness binary particle swarm optimization, such as lack of global search ability, easy to fall into local optimum and slow convergence speed, a new adaptive parameter strategy and a particle divergence index are proposed, which are combined with simulated annealing mechanism to improve the optimization ability of the algorithm. In order to test the performance of the algorithm, simulation experiments are carried out by selecting the knapsack problem case library with different dimensions and the UCI feature selection problem case library with different scales, and the experimental data are statistically analyzed. The experimental and analytical results show that the proposed algorithm is superior to the comparison algorithm in optimization accuracy, algorithm stability and convergence speed.