In this paper, a fast-sorting method called summation of normalized objectives and diversified selection (SNOV-DS) is embedded in Comprehensive Learning Particle Swarm Optimization (CLPSO) to solve multi-objective problems. Due to this method, the simulation time will be decreased. The convergence to true Pareto front and the spread of solutions can also be improved. The algorithm is tested on a set of commonly used multi-objective benchmark functions. The simulation results show that the proposed algorithm is competitive in terms of both performance and running speed.