To avoid the premature convergence and enhance the search capability of the high-dimensional space, a novel self-perception high-dimensional chaotic particle swarm algorithm was presented. Firstly, a double perturbation of pBest and gBest was used to enhance the searching capability of particles. Secondly, self-perception approach was proposed to help the particle swarm to avoid the premature convergence. Lastly, three discrete PSO variants were tested on the traveling salesman problem (TSP). Experimental results show that the self-perception high-dimensional chaotic particle swarm algorithm is simple, effective and promoting in a high-dimensional space.