Abstract: Research on workflow service subject selection issues, in workflow management system, work load ability difference, and the system load is dynamic, the traditional algorithm is difficult to obtain the optimal workflow service subject optimization scheme, leading to low utilization rate of system resources.In order to keep the system responsible for balance, raise the utilization rate of resources system, a method based on particle swarm algorithm optimization method of workflow service subject.The workflow services body preferably built corresponding mathematical model, and then the particle swarm optimization algorithm to solve the model, workflow service main body the optimal solution, finally carries on the simulation test.The test results show that, compared with the traditional method, particle swarm optimization algorithm can be used for different types of task allocation in different working machine, improve the utilization rate of system resources, to achieve a variety of resource load balancing system.