To improve the radiation measurement efficiency of nuclear retirement facilities and reduce the risk of radiation exposure to measurement personnel, a radiation patrol control system for multiple unmanned vehicle formations has been designed. Firstly, the navigation following formation strategy is adopted to control the robots to move in a predetermined formation, while collecting real-time radiation intensity information and their respective position data measured by each unmanned vehicle during the formation process, to preliminarily analyze the radiation distribution inside the environment. Secondly, utilizing radiation intensity and location information, the Markov chain Monte Carlo method is employed to estimate the parameters of the radiation source. The simulation results show that the unmanned vehicle formation can move along the automatically planned path in radiation environment, with advantages such as fast response speed, high control accuracy, and it can estimate the parameters of the radiation source position coordinates.