Classical complex networks mainly describe same type of entities and one type of interrelations between the entities. Multi-subnet composited network is a model that describes different types of entities and multiple types of interrelations between the entities. Dynamic reorganization of this model provide two operations: Compounding (combine two subnets into a ‘bigger’ one) and reducing (obtain a ‘small’ network from a ‘big’ one). In this paper, a vector-composited network is defined by importing multi-dimensional space, which converts the interrelations between entities into multi-dimensional vector. Dynamic reorganization of networks is converted into base transformations in multi-dimensional space. Formalized descriptions of compounding and reducing are presented. Further, vector-composited network of passenger transport with high speed and low speed railways in mainland China is established by empirical data. Topological analysis of networks obtained by dynamic reorganizations illustrates the development of railway system in mainland China.