Based on the characteristics of multi-resource leveling in multiple projects scheduling problem, a multi-objectives
optimization model is setup. By applying Pareto optimal method into vector evaluated particle swarm optimization (VEPSO),a new vector evaluated particle swarm optimization based on Pareto (VEPSO-BP) is proposed. Finally, the performance of VEPSO-BP is tested with a testing example which is compared with VEPSO. Experiment results show that, VEPSO-BP is better than VEPSO in convergence efficiency, which also performs well in solving multi-resource leveling in multiple projects scheduling problem.