In order to solve the problem of agent imbalance caused by the traditional contact center serving different types of customers, improve the service quality of the contact center and optimize customer service experience, some multi skilled agents were introduced. Without increasing the number of agents, a part of the multi skilled agents were transferred in real time according to the queue length and agent status. A new contact center model which aimed to minimize customer abandonment was constructed with the goal of the lowest customer abandonment rate. Arena was used to build a model to simulate and analyze the overall performance of the new contact center model and compared with the traditional model to verify its effectiveness. It is found that the new contact center model not only improves the connection rate, but also reduces the waiting time of customers and balances the service intensity of agents.