Abstract: This paper analyzed and established mathematic models for vehicle routine p roblem under the con2 dition of stock shortage. To solve the models, it p resented a dynamic particle swarm op timization algorithm based on sub2group collaboration. Finally, the paper made some experimental calculations, and the results of calculations p roved that the algorithm could avoid blind search effectively, and overcome the limitation of easily trapp ing in local extreme points and leading to p remature, as a result, it had better capability of global op timization, higher speed of convergence and p recision than standard particle swarm op timization.