In this paper, we propose a mobile robot formation control method and an efficient formation transformation strategy for an engineering problem where mobile robot formation transformation is inflexible in a complex obstacle environment. First, kinematic and positional models of differentially mobile robots are designed for accurate control of robot formation; Second, an improved leader-follower formation method for introducing virtual robots is designed, and the data set for robot formation avoidance strategy is proposed by considering the obstacle data in the environment comprehensively, which lays the foundation for dynamic robot formation transformation； Finally, a formation strategy with a robot formation scaling factor and a robot formation dynamic transformation factor  is designed. This strategy can quickly determine the position of each robot in the formation through the parameter matrix of the target formation based on the information of obstacles in the environment detected by the robot, thus realizing the dynamic transformation of the robot formation. We have verified its effectiveness through computer simulation and physical experiments. The strategy of this paper realizes the robot formation through the complex obstacle environment, and successfully realizes the functions of formation maintenance, isomorphic deformation, heterogeneous deformation and scattered obstacles crossing obstacles.